Assessment of Out-of-Pocket Spending for COVID-19 Hospitalizations in the US in 2020

This cross-sectional study assesses out-of-pocket spending by privately insured and Medicare Advantage patients for COVD-19 hospitalizations in 2020.


Introduction
From August 2020 through July 2021, there were 2.4 million US hospitalizations for COVID-19. 1 To mitigate patient financial burden, many private insurers and Medicaid Advantage insurers voluntarily waived cost sharing for COVID-19 hospitalizations during part or all of 2020. 2,3 The literature examining cost sharing for other respiratory infection-related hospitalizations suggests that these waivers potentially resulted in substantial savings for patients. [4][5][6] For example, among privately insured patients hospitalized for treatment of respiratory infections between 2016 and 2019, average out-of-pocket spending was $1653 for those in traditional plans and $1961 for those in consumer-driven health plans. 4 Among Medicare Advantage patients hospitalized for treatment of influenza in 2018, mean out-of-pocket spending was almost $1000. 6 Although waivers may have mitigated the financial burden for many patients hospitalized for treatment of COVID-19 during 2020, some patients may still have been billed if their plans did not implement waivers or if waivers did not capture all hospitalization-related care. Hospitalizations can result in 2 categories of bills. 7,8 The first includes facility services provided by hospitals, such as accommodation and inpatient pharmacy services. The second includes services from clinicians and ancillary service providers (hereafter referred to as professional and ancillary services). This category includes clinician services for emergency department and inpatient care as well as ambulance services for transport to the hospital. Although waivers would ideally cover both categories, some may have covered only facility services billed by hospitals, not professional and ancillary services billed separately by professionals providing those services.
Although protecting patients from the costs of hospitalization is an important goal regardless of condition, protecting patients from the costs of COVID-19 hospitalizations specifically may be especially important given the number of hospitalizations that may occur and given that the threat of cost sharing could deter patients with serious COVID-19 symptoms from seeking care. Despite this, to our knowledge, no study has assessed the amount for which patients were billed for COVID -19 hospitalizations during 2020 either overall or by service category. Addressing this knowledge gap may inform policy in several ways. First, it may demonstrate the potential financial burden patients may experience if insurers allow cost-sharing waivers to expire, as many chose to do during 2021. 9,10 Second, it may motivate efforts to improve the comprehensiveness and implementation of the remaining insurer cost-sharing waivers for COVID-19 hospitalizations. Third, it may indicate the potential need for federal legislation mandating US insurers to waive cost sharing for these hospitalizations; this legislation was proposed but not passed in the US House of Representatives in 2020. 11 Fourth, it may inform cost-sharing policies for hospitalizations during future pandemics. In this study, we used national claims data to estimate out-of-pocket spending for COVID-19 hospitalizations from March to September 2020 among patients covered by private insurance and Medicare Advantage plans.

Data Source
In May 2021, we conducted a cross-sectional analysis of the IQVIA PharMetrics Plus for Academics database (IQVIA Inc). This database contains fully adjudicated medical and pharmacy claims from deidentified patients in all 50 states and the District of Columbia. Claims were complete through

Study Sample
We included hospitalizations that had a confirmed primary diagnosis of COVID-19 infection (ICD-10-CM diagnosis code U071) that began and ended between March 1 and September 29, 2020. We required discharge before September 30, 2020, to ensure that the end of hospitalization was observed (eAppendix 1 in the Supplement gives the details). We excluded hospitalizations if they were covered by a secondary insurer (eg, a different private insurance plan) or if any associated claim had missing data for out-of-pocket spending or billing provider type.

Categorization of Claims
For each hospitalization, we assigned claims with the corresponding hospitalization identifier to 1 of 3 mutually exclusive categories: facility services, professional and ancillary services, and unclassified services (eAppendix 2 in the Supplement gives details). Claims for facility services were defined as institutional claims with a hospital or emergency department place of service and a hospital billing provider type. These services included but were not limited to hospital accommodation and inpatient laboratory and pharmacy services. Claims for professional and ancillary services were defined as 1 of 3 types of services: ambulance (claims with an ambulance place of service or procedure code), clinician (claims with an emergency department or hospital place of service and clinician billing provider type), and miscellaneous (claims with billing provider type for miscellaneous providers, such as durable medical equipment providers). For additional context, clinician services were divided into 4 subtypes: emergency department (claims with an emergency department place of service), inpatient evaluation and management (claims with a hospital place of service and procedure code for evaluation and management, such as initial hospital care), inpatient diagnostic testing (claims with a hospital place of service and procedure codes for laboratory tests, radiology tests, electrocardiography, echocardiography, electroencephalography, and vascular diagnostic studies), and other inpatient services (claims with hospital place of service and procedure codes for services other than evaluation and management and diagnostic testing, such as procedures). Claims for unclassified services were the 4.3% of claims that were assigned the confinement identifier for the COVID-19 hospitalization but did not meet criteria for a facility or professional or ancillary service. For three-quarters of these claims, the place of service was office, home, or hospital outpatient department. Although some could represent care provided at visits resulting in direct hospital admission, others could represent care provided at unrelated visits. In the main analysis, we excluded these claims to maximize the probability of capturing only out-of-pocket spending for services truly associated with hospitalizations. These claims were included in a sensitivity analysis (eAppendix 3 in the Supplement).

Outcomes
Out-of-pocket spending was the sum of deductibles, co-insurance, and co-payments; this quantity excluded any surprise bills for out-of-network care. 8 For each payer type (private insurance and Medicare Advantage), we determined the proportion of hospitalizations in 2 categories: those that had out-of-pocket spending for facility services (with or without out-of-pocket spending for professional and ancillary services) and those that had out-of-pocket spending for facility services, professional and ancillary services, or both. For hospitalizations in both categories, we calculated total out-of-pocket spending, defined as the sum of out-of-pocket spending across facility and professional and ancillary services. In addition, we calculated the proportion of all hospitalizations with out-of-pocket spending for the 3 main types of professional and ancillary services and for the 4 subtypes of clinician services.

Presence of Cost-Sharing Waivers
The database did not report whether COVID-19 hospitalizations were covered by plans with costsharing waivers. However, as noted in the Results section, few hospitalizations in our sample had cost sharing for facility services. Although this might suggest that most hospitalizations were covered by insurers that waived cost sharing for facility services (ie, that the absence of cost sharing for facility services implied the presence of a waiver), a potential alternative explanation is that most patients had already met their plan's annual out-of-pocket maximum at the time of the hospitalization. To evaluate this possibility, we restricted analyses to hospitalizations of patients continuously enrolled since January 2020, calculated out-of-pocket spending across medical and pharmacy claims in 2020 before the hospitalization, and calculated the incidence of out-of-pocket spending for facility services among hospitalizations for patients in the lowest quartile of this prior out-of-pocket spending. These patients likely had not met out-of-pocket maximums at the time of their hospitalization. If few of these patients had cost sharing for facility services, cost-sharing waivers, rather than meeting out-ofpocket maximums, may have been associated with the low observed incidence of cost sharing for facility services.
We also explored whether it was reasonable to assume that hospitalizations with out-of-pocket spending for facility services were not covered by insurers with cost-sharing waivers for these services (ie, that the presence of cost sharing for facility services implied the absence of a waiver-the inverse of the previously mentioned assumption). To evaluate this assumption, we compared the incidence of out-of-pocket spending for facility services between hospitalizations for COVID-19 and those for influenza. The latter require care similar to that required by COVID-19 hospitalizations, but to our knowledge, no insurers waived cost sharing for influenza hospitalizations during the study period. If the presence of out-of-pocket spending for facility services implies the absence of a waiver for these services, a higher proportion of influenza hospitalizations would have out-of-pocket spending for facility services compared with COVID-19 hospitalizations. In this analysis, influenza hospitalizations were those that met similar inclusion and exclusion criteria but had a primary diagnosis of influenza (ICD-10-CM diagnosis code J09-J11). None of the influenza hospitalizations included had claims with a COVID-19 diagnosis code (U017).

Statistical Analysis
We used descriptive statistics to assess patient characteristics, length of hospital stay, and intensive care unit use (eAppendix 2 in the Supplement). To contextualize cost-sharing amounts, we calculated mean and median allowed amounts (reimbursement to providers plus patient liability) across facility and professional and ancillary services among privately insured and Medicare Advantage hospitalizations separately. Analyses were performed using SAS, version 9.4 (SAS Institute Inc).

Out-of-Pocket Spending
Of  Of all 2698 hospitalizations for Medicare Advantage patients, 7 (0.3%) had total out-of-pocket spending greater than $2000 and 5 (0.2%) had total out-of-pocket spending greater than $4000.

Analyses Assessing Presence of Cost-Sharing Waivers
Among hospitalizations for privately insured and Medicare Advantage patients in the lowest quartile of out-of-pocket spending before hospitalization, the proportion with out-of-pocket spending for facility services was 8.3% for privately insured patients and 1.8% for Medicare Advantage patients  (Table 2). In the sensitivity analysis including claims for unclassified services that did not meet the criteria for a facility or professional or ancillary service, the results were not substantially different from those of the main analysis.

Discussion
In

Strengths and Limitations
This study has strengths. We used a national database that includes both privately insured and Medicare Advantage plans. These plans are important sources of coverage for adults aged 50 and older, a group that is at high risk for COVID-19 hospitalization. [16][17][18] This study also has limitations. First, we cannot prove that COVID-19 hospitalizations in this study were mostly covered by plans with cost-sharing waivers. Second, if patients did not pay the amounts they were billed or were not billed because they died in the hospital, the incidence of actual out-of-pocket spending would differ from the incidence estimated by this study. However, the amount billed to patients still shows the financial burden patients may experience without costsharing waivers. Third, the number of hospitalizations with out-of-pocket spending for facility services was small, likely owing to the widespread presence of insurer cost-sharing waivers during 2020. Consequently, mean total out-of-pocket spending among these hospitalizations may be imprecisely estimated. Fourth, our sample of 4075 COVID-19 hospitalizations represents a small proportion of the roughly 311 000 hospitalizations in the US from March to September 2020. 19 Thus, results may not necessarily be generalizable to all privately insured and Medicare Advantage patients. However, most hospitalizations of privately insured patients in our study were covered by preferred provider organization plans, and most hospitalizations of Medicare Advantage patients were covered by health maintenance organizations, consistent with the national distribution of plan types among privately insured and Medicare Advantage enrollees. 16,20 Fifth, findings on out-ofpocket spending may not be generalizable to traditional Medicare enrollees. For lengthy hospitalizations, such as those for patients with COVID-19 infection, cost sharing is typically lower for traditional Medicare enrollees compared with Medicare Advantage enrollees. 21

Conclusions
The findings of this cross-sectional study suggest that insurer cost-sharing waivers for COVID-19 hospitalizations may not always capture all hospitalization-related care. Moreover, patient financial burden for COVID-19 hospitalizations could be substantial without insurer waivers. The increasing trend toward abandonment of these waivers suggests that relying on voluntary actions by insurers is not an ideal strategy if policy makers wish to protect patients from the costs of COVID-19 hospitalizations. 9,10 To achieve this goal, federal policy makers might consider legislation mandating insurers to waive cost sharing for COVID-19 hospitalizations throughout the public health emergency. 11 Such a mandate would ideally include all hospitalization-related care, similar to existing federal mandates that require insurers to fully cover all direct and related costs of COVID-19 testing and vaccines. 13 Future research should include monitoring of patient financial burden resulting from COVID-19 hospitalizations as coverage policies change.