Economic burden of sickle cell disease in the United States: a retrospective analysis of a commercial insurance database

Sickle cell disease (SCD) is the most common inherited hemoglobinopathy in the US [1]. Based on records from universal newborn SCD screening programs that have been implemented across the US since 2006, the crude birth prevalence for SCD between 2016 and 2020 has been estimated to be one in every 2070 live births [2,3]. Among the estimated 100,000 individuals in the US affected by SCD, approximately 60% are adults ≥18 years of age [1,4]. Mutations to the β-globin gene and the ensuing formation of sickle hemoglobin (HbS) result in the sickling of erythrocytes that is characteristic of SCD. Through hemolysis, vaso-occlusive crisis (VOC) or other property changes in the red blood cell, the complications of this lifelong illness affect multiple organ systems, leading to substantial morbidity and increased mortality [5–7].

In the US, the medical costs of patients with SCD are substantial due to the healthcare services required for the management of SCD and its complications, with inpatient care accounting for a significant portion of the costs [8]. Among pediatric patients with SCD, inpatient hospitalizations have been reported to cost up to $900 million per year [9]. Moreover, the lifetime medical costs associated with SCD have been estimated to be $1.7 million per patient [10]. While the healthcare resource utilization (HCRU) and costs of SCD have been assessed in a number of published studies, most of these analyses were conducted among patients with noncommercial insurance provided by Medicaid and Medicare, government-funded insurance plans that provide healthcare coverage for low-income individuals and those ≥65 years of age or <65 years with certain disabilities or conditions, respectively [11]. Although a handful of studies have described the HCRU and costs of commercially insured SCD patients with healthcare coverage purchased from private companies, their focus was largely on patients with severe SCD as evidenced by VOCs and/or other SCD-related complications [12–14]. Three studies investigated the burden of SCD among commercially insured patients with and without SCD [10,15,16]. The first study evaluated the burden of SCD among patients who had recurrent VOCs [16]. While the second study evaluated HCRU stratified by service category (e.g., inpatient, outpatient and office visits, etc.), the SCD and control cohorts were constructed using two different data sources and estimates of SCD-attributable costs were limited to the total medical costs paid by insurers and out-of-pocket (OOP) costs paid by patients, which could include copays, deductibles, coinsurance and any medical care or supplies not covered by insurers; the corresponding healthcare service costs were not reported [10]. The third study was a continuation of a prior study that provided a granular description of the HCRU and costs of pediatric patients with SCD [12]. However, the follow-up study only evaluated the economic burden of pediatric SCD [15]. Furthermore, this study was conducted using data for healthcare services received nearly two decades ago, as did several of the other studies of commercially insured populations [10,12,13].

There is limited information available regarding the contemporary real-world clinical and financial burden of patients with SCD in the US who have commercial healthcare coverage. In particular, the medical costs of specific healthcare service categories as well as the OOP costs of patients with SCD are largely unknown. The aim of this study was to evaluate the all-cause HCRU and healthcare costs of commercially insured pediatric and adult patients with SCD in the US versus matched control patients without SCD in a more contemporary dataset.

In this retrospective cohort study of pediatric and adult patients with SCD and commercial healthcare coverage in the US, both HCRU and costs were substantially higher among patients with SCD compared with their matched non-SCD counterparts. The overall costs for the SCD cohort were 11-fold and 6-fold that of their matched controls for pediatric and adult patients, respectively. Hospitalizations accounted for the largest proportion of the total healthcare costs for both pediatric and adult patients with SCD. Compared with patients without SCD, the hospitalization costs of those with SCD were 32- to 35-fold higher for pediatric patients and 15- to 17-fold higher for adult patients. ER visits accounted for the next highest fold difference with seven- to eightfold higher costs for both pediatric and adult patients with SCD. Correspondingly, hospitalizations and ER visits accounted the most differences in HCRU compared with the non-SCD patients.

These findings are consistent with the few previously published analyses of commercially insured patients with SCD. HCRU and healthcare costs were evaluated among 621 commercially insured pediatric patients (aged 1–17 years) with SCD in 2005 [12]. In comparison to our study, similar proportions of pediatric patients with SCD were hospitalized (38 vs 33%, mean of 0.8 vs 0.6 admissions per patient), had ER visits (45 vs 50%, mean of 0.9 vs 1.2 visits per patient), and outpatient pharmacy utilization (82 vs 80%, mean of 10.8 vs 12.8 pharmacy claims per patient). However, the pediatric SCD patients with at least one hospitalization from the previous study had a shorter mean total hospital LOS than those in our study (3.9 vs 7.3 days). In comparison to pediatric patients with SCD, the mean total annual healthcare costs of the 2,550,959 pediatric patients without SCD were 11-fold lower ($14,722 vs $1293) [12,15]. When stratified by resource category, pediatric patients with SCD had 34-fold higher inpatient costs, eightfold higher outpatient costs, and twofold higher pharmacy costs compared with pediatric patients without SCD [15]. After taking into account inflation rate adjustment from 2005 to 2020, the costs in our study are higher than those of the previous study. In part, this may be due to study duration, where our study includes healthcare services incurred between 2016 and 2020 while the other study covers a single year, 2005. A longitudinal study covering 5-year periods between 2010 and 2018 reported annual inpatient and ER admission rates that varied by 19.8 and 15.8%, respectively [13]. In part, the higher costs of our study may also be due to pediatric patients who received an HSCT or bone marrow transplant in the follow-up period. While the costs of transplantation were not evaluated in this study, the total costs of allogeneic HSCT at 1 year have been reported to range from $69,218 to $637,193 [20]. Among the pediatric patients with SCD in our study, total costs at the 97.5th percentile were $182,106, potentially representing transplant-related costs. However, while mean costs may have been skewed by the high costs of HSCT, the median total costs of patients with SCD were still substantially higher than that of their matched non-SCD counterparts ($9042 vs $817, 11.1× fold difference). In addition, one of the largest discrepancies was the proportion of the total costs due to pharmacy usage; pharmacy usage accounted for a greater proportion of the total costs in our study (11.1 vs 3.6%). The higher costs could potentially be due to the increased use of hydroxyurea (35 vs 11%), pain medication (69 vs 45%) and iron chelation agents (5.6 vs 0.2%) in our study [12].

In a prior study by Johnson et al. [10] where 20,891 patients were followed from 2007 to 2018, the lifetime burden of total SCD-related medical costs among non-elderly patients with commercial insurance was estimated to be $1.7 million (2020 USD) with $44,000 in OOP costs. Compared with their matched non-SCD counterparts, the mean number of hospitalizations (0.57 vs 0.04), ER visits (1.3 vs 0.14), office visits (4.6 vs 2.4) and prescription fills (8.0 vs 1.6) per year were higher among the pediatric patients (aged ≤18 years). Patients 19–30 and 31–64 years of age demonstrated the same trend, with higher utilization among those with SCD (age 19–30: 0.69 vs 0.07 hospitalizations, 1.9 vs 0.2 ER visits, 4.5 vs 2.4 office visits and 9.8 vs 2.4 prescription fills; age 31–64: 0.46 vs 0.07 hospitalizations, 1.12 vs 0.18 ER visits, 6.5 vs 3.7 office visits and 15.9 vs 8.5 prescription fills). Similar to our study, total medical costs, OOP costs and the utilization of corresponding resource categories were higher among patients with SCD and increased with age.

A longitudinal analysis of the economic burden of severe SCD assessed all-cause HCRU and costs over a 5-year period among patients with commercial and Medicare coverage from 2010 to 2018 [13]. Among the overall population, 48.4–68.2% of patients were hospitalized each year, with a mean (SD) of 1.2 (2.2) to 1.7 (2.3) admissions per year; the mean (SD) LOS per admission was 4.1 (3.7) days. Annually, 57.6–73.4% of patients had an ER visit, with a mean (SD) of 2.1 (5.3) to 2.6 (4.4) visits per year. In our study, the proportions of patients who were hospitalized (50.1–53.4%) and had an ER visit (69.8–70.2%) were also comparable to that of the longitudinal study, as were the mean number of hospitalizations PPPY during the overall follow-up period (1.1–1.5) among patients who were hospitalized. However, hospital LOS was longer (4.8–8.1 days) and there were fewer ER visits (1.5–2.2 visits per year). When stratified by age, the pediatric patients (<18 years of age) in the longitudinal study had mean annual costs (2018 USD) of approximately $29,250 for hospitalizations, $2210 for ER visits, $1380 for office visits and $4000 for outpatient pharmacy usage, resulting in total costs of $52,210. Compared with these pediatric patients, costs were generally higher among patients 18–30 years of age while patients ≥31 years had intermediate costs. Apart from office visits, all other resource categories reported in the longitudinal study had higher utilization and costs compared with our study. In part, this may be due to differences in the composition of the study populations. Patients in the longitudinal study were required to have at least one VOC within the first 2 years of follow-up and thus might have been selected as higher-risk patients than those in our study. Yet the two studies showed many comparable results and in our study population, 71% of the pediatric patients and 69% of the adult patients had at least one VOC during the entire follow-up period (median of 34 and 29 months, respectively).

This finding of higher costs among patients with VOCs is supported by two studies that evaluated the impact of VOCs on the healthcare costs of patients with SCD [14,16]. In one study of 16,092 commercially insured patients ≥16 years of age from 2001 to 2017 who had 0, 1 or ≥2 VOC episodes within a 1-year follow-up period, mean annual total all-cause healthcare costs increased with the number of VOCs from $15,747 (2018 USD, no VOCs) to $27,194 (1 VOC) and to $64,555 (≥2 VOCs) [14]. While inpatient costs increased threefold for patients with 0–1 VOC and 2.9-fold for those with 1 to ≥2 VOCs, the corresponding increases for outpatient costs were lower at 1.2- and 1.5-fold; ER costs increased by 1.9- and 2.6-fold. Another study found that recurrent VOCs, defined as ≥2 VOCs per year for 2 consecutive years, resulted in even higher costs [16].

Similar to our study, these previously published studies utilized data from the MarketScan Commercial database for their analyses, allowing for a crude assessment of SCD-related HCRU and costs over time. Unlike the life expectancy of patients with SCD, which has trended higher from 38 to 42 years in 2005 to 54 years in 2019, the economic burden of SCD has not improved significantly over the same time period [7,21]. The substantial HCRU and costs of patients with SCD continue to be primarily due to a greater need for inpatient and ER admissions to manage complications such as VOCs and other SCD-related chronic complications. This is likely due to the limited treatment options available for the management of SCD-related complications. Prior to 2017, hydroxyurea was the only FDA-approved pharmacological intervention available for the treatment of SCD. Studies have demonstrated that hydroxyurea treatment is associated with decreased VOC events, fewer hospitalizations and ER admissions, and lower healthcare costs [22–24]. However, as demonstrated in our study, hydroxyurea treatment remains underutilized [25]. So, there remains a gap to improve healthcare treatments for SCD patients. In addition, previous studies reported that approximately 45% of commercially insured patients with SCD were prescribed pain medication [12,14]. Among patients with recurrent VOCs or other complications indicative of severe SCD, the proportion of patients prescribed an opioid increased to over 90% [13,16]. In our study, 69% of the pediatric and 86% of the adult populations required pain medication during the overall follow-up period; 61 and 78% of the pediatric and adult patients, respectively, were prescribed an opioid. This widespread use of pain medication further demonstrates that there is an unmet need for treatments to improve the management of SCD and its complications and reduce the associated economic burden.

In our study, 58 and 56% of the pediatric and adult patients with SCD, respectively, had PPO or point-of-service coverage. PPO and point-of-service healthcare plans enhance access to the specialist care needed for the treatment of SCD by permitting the use of out-of-network healthcare providers, although this comes with increased OOP costs for patients. However, nearly 20% of patients had HMO or exclusive provider organization healthcare plans, which do not offer coverage for out-of-network providers. As a result, patients may not seek required medical care if it cannot be obtained through in-network providers. Moreover, care obtained through out-of-network providers may not be reliably captured in the database, resulting in possible underestimation of the HCRU and costs of these patients. Further research into the potential impact of different levels of commercial insurance coverage on the management of SCD and its complications is warranted. In addition, it has been demonstrated that SCD healthcare disparities are considerable compared with other orphan diseases, resulting in poorer outcomes [26]. Countries with a higher prevalence of SCD, such as countries in sub-Saharan Africa and India, have implemented comprehensive programs to improve the management of SCD and reduce disease burden; these programs include early screening, public education to increase disease awareness and increased healthcare access [27]. Such programs may be effective in reducing the clinical and economic burden of SCD [28]. However, additional research is required to evaluate the benefit of such preventive programs on the economic burden of SCD among commercially insured patients with SCD.

The findings of this study should be interpreted in the context of certain limitations. The administrative claims included in this database are collected for billing purposes. As a result, these records lack clinical detail and may be subject to miscoding or coding errors. Since the MarketScan Commercial database does not report race, patients could not be matched on that characteristic. Furthermore, the Commercial database is not representative of all privately insured patients but only those covered by certain employer-sponsored health plans. In addition, the clinical manifestation of SCD varies based on the inherited hemoglobin variants; hemoglobin SC (HbSC) and hemoglobin SS (HbSS) result in clinically milder and more severe forms of SCD, respectively. In our study, no distinction was made between patients with SCD genotypes that would require different levels of care, such as HbSC and HbSS, since genotype information may not be captured for all individuals in the database. However, the adult SCD population may have lower disease severity since they were able to maintain employment. Accordingly, SCD-related HCRU and costs may be underestimated. Moreover, healthcare during 2020 may have been underutilized by both the patients with SCD and their matched non-SCD counterparts due to the COVID-19 pandemic. Although the database is a very large convenience sample, our study findings may not generalize to the entire US or regions not well represented in the data source, as well as to patients who are uninsured or have other types of healthcare coverage. Finally, this study excluded individuals without commercial insurance, who may have a greater economic burden.

This retrospective cohort analysis highlights the substantial economic burden of SCD among patients enrolled in US commercial insurance plans. Both the pediatric and adult populations had substantially higher HCRU compared with their matched controls. Furthermore, costs for payers as well as OOP costs for patients were higher across all resource categories. These findings align with previous studies reflecting the burden of illness associated with SCD. They also suggest that more optimal SCD treatments may have the potential to reduce the economic burden for patients with SCD, as well as payers.

All authors were responsible for study conception and design, acquisition of data, data analysis and interpretation, and drafting and revision of the manuscript.

The authors acknowledge G Lin, an employee of Novosys Health, who provided editorial support in the drafting of this manuscript.

This study was funded by Pfizer, Inc. Novosys Health received financial support from Pfizer Inc., in connection with the development of this manuscript. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

GT Barcelos is a current employee and stockholder of Pfizer AG; T Peixoto was an employee of Pfizer Portugal at the time the study was conducted; CL Baker is a current employee and stockholder of Pfizer Inc.; J Alvir was an employee of Pfizer Inc. at the time the study was conducted; J Lin is a current employee of Novosys Health LLC, which received research funding from Pfizer Inc. The authors have no other competing interests or relevant affiliations with any organization or entity with the subject matter or materials discussed in the manuscript apart from those disclosed.

Medical writing assistance was provided by G Lin of Novosys Health, and was funded by Pfizer, Inc.

This study utilized de-identified, HIPAA-compliant patient data from the IBM MarketScan Commercial Database; institutional review board approval was not required.

This study used data available from IBM’s MarketScan Commercial Database. Restrictions apply to the availability of these data, which were used under a licensing agreement.

This work is licensed under the Attribution-NonCommercial-NoDerivatives 4.0 Unported License. To view a copy of this license, visit https://creativecommons.org/licenses/by-nc-nd/4.0/