key: cord-0717686-y0obg5q5
authors: Chua, K.-P.; Conti, R. M.; Becker, N. V.
title: Out-of-Pocket Spending for COVID-19 Hospitalizations in 2020
date: 2021-05-30
journal: nan
DOI: 10.1101/2021.05.26.21257879
sha: 6eff63f057d34c382986f38c165862054d12af8d
doc_id: 717686
cord_uid: y0obg5q5

IMPORTANCE: Many insurers waived cost-sharing for COVID-19 hospitalizations during 2020. Nonetheless, patients may have been billed if their plans did not implement waivers or if waivers did not capture all hospitalization-related care, including clinician services. Assessing out-of-pocket spending for COVID-19 hospitalizations in 2020 could demonstrate the financial burden patients may face if insurers allow waivers to expire, as many chose to do during early 2021. OBJECTIVE: To estimate out-of-pocket spending for COVID-19 hospitalizations in 2020 DESIGN: Cross-sectional analysis SETTING: IQVIA PharMetrics Plus for Academics Database, a national claims database PARTICIPANTS: COVID-19 hospitalizations for privately insured and Medicare Advantage patients during March-September 2020 MAIN OUTCOMES/MEASURES: Mean total out-of-pocket spending, defined as the sum of out-of-pocket spending for facility services billed by hospitals (e.g., accommodation charges) and for professional/ancillary services billed by clinicians and ancillary providers (e.g., clinician inpatient evaluation and management, ambulance transport) RESULTS: Analyses included 4,075 hospitalizations. Of the 1,377 hospitalizations for privately insured patients and the 2,698 hospitalizations for Medicare Advantage patients, 981 (71.2%) and 1,324 (49.1%) had out-of-pocket spending for facility services, professional/ancillary services, or both. Among these hospitalizations, mean (SD) total out-of-pocket spending was $788 (1,411) and $277 (363). In contrast, 63 (4.6%) and 36 (1.3%) hospitalizations had out-of-pocket spending for facility services. Among these hospitalizations, mean total out-of-pocket spending was $3,840 (3,186) and $1,536 (1,402). Total out-of-pocket spending exceeded $4,000 for 2.5% of privately insured hospitalizations, compared with 0.2% of Medicare Advantage hospitalizations. CONCLUSIONS: Few COVID-19 hospitalizations in this study had out-of-pocket spending for facility services, suggesting most were covered by insurers with cost-sharing waivers. However, many hospitalizations had out-of-pocket spending for professional/ancillary services. Overall, 7 in 10 privately insured hospitalizations and half of Medicare Advantage hospitalizations had any out-of-pocket spending. Findings suggest insurer cost-sharing waivers may not cover all hospitalization-related care. Moreover, high cost-sharing for some hospitalizations suggests out-of-pocket burden could be substantial if waivers expire, particularly for privately insured patients. Rather than rely on voluntary insurer actions to mitigate this burden, federal policymakers should consider mandating insurers to waive cost-sharing for all COVID-19 hospitalization-related care throughout the pandemic.

to our knowledge has assessed the amount patients were billed for COVID-19 hospitalizations 1 0 1 during 2020, either overall or by service category. Addressing this knowledge gap may inform 1 0 2 policy in several ways. First, it could motivate efforts to improve the comprehensiveness and sharing waivers to expire, as many chose to do during early 2021. 9,10 Finally, it could illustrate 1 0 6 the potential need for federal legislation mandating U.S. insurers to waive cost-sharing for these 1 0 7

hospitalizations -legislation that was proposed, but not passed, in the U.S. House of 1 0 8

Representatives in 2020. 11 In this study, we used national claims data to estimate out-of-pocket plans in 2020. The database did not include any patients covered by self-insured private plans. Modification (ICD-10-CM) diagnosis codes, a hospitalization identifier assigned to all claims 1 2 3 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted May 30, 2021. ;  that occurred on or between the admission and discharge dates of hospitalizations, amounts 1 2 4 billed to patients (deductibles, co-insurance, and co-payments), and a flag for whether the billing 1 2 5 provider was a hospital, clinician, or other entity. The database does not include information on 1 2 6 race, ethnicity, household income, out-of-network status, or in-hospital death (to protect patient 1 2 7 confidentiality). Moreover, the database does not include plan identifiers or information on plan 1 2 8 benefit design, including whether insurers had cost-sharing waivers for COVID-19 1 2 9

hospitalizations. As discussed below, we conducted analyses to evaluate whether such waivers 1 3 0 may have been in place. Because data were de-identified, the Institutional Review Board of the COVID-19 infection (ICD-10-CM diagnosis code U071) and that began and ended between 1 3 4

March 1-September 29, 2020. We required discharge before September 30, 2020 to ensure the 1 3 5 end of hospitalizations was observed (see Appendix 1 for details). We excluded hospitalizations 1 3 6 if they were covered by a secondary insurer (e.g., a different private insurance plan) or if any 1 3 7 associated claim had missing data for out-of-pocket spending or billing provider type. place of service and a hospital billing provider type). These services included but were not 2) Claims for professional/ancillary services, defined as one of three types of services: 1 4 6 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted May 30, 2021. ;

• Ambulance (claims with an ambulance place of service or procedure code) 1 4 7

• Clinician (claims with an emergency department or hospital place of service and clinician 1 4 8 billing provider type) 1 4 9

• Miscellaneous (claims with billing provider type for miscellaneous providers, such as 1 5 0 durable medical equipment providers and dialysis centers). For additional context, clinician services were divided into 4 subtypes:

• Emergency department (claims with an emergency department place of service) 1 5 3

• Inpatient evaluation and management (claims with a hospital place of service and 1 5 4 procedure code for evaluation and management, e.g., initial or subsequent hospital care) 1 5 5

• Inpatient diagnostic testing (claims with hospital place of service and procedure codes for 3) Unclassified claims. This category included the approximately 4.3% of claims that were 1 6 2 assigned the confinement identifier for the COVID-19 hospitalization but did not meet criteria 1 6 3 for a facility or professional/ancillary service. Three-quarters of these claims had a place of 1 6 4 service code for office, home, or hospital outpatient department. While some of these claims 1 6 5 could represent care at visits resulting in direct admission to the hospital, they could also include 1 6 6 care provided at unrelated visits. In the main analysis, we excluded these claims to maximize the probability of only capturing out-of-pocket spending for services truly associated with 1 6 8 hospitalizations. We included these claims in a sensitivity analysis. . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted May 30, 2021. ;

Outcomes. Out-of-pocket spending was defined as the sum of deductibles, co-insurance, 1 7 0 and co-payments. For each payer type (private insurance and Medicare Advantage), we 1 7 1 determined the proportion of hospitalizations in two categories: those with out-of-pocket 1 7 2 spending for facility services (with or without out-of-pocket spending for professional/ancillary 1 7 3 services), and those with out-of-pocket spending for facility services, professional/ancillary 1 7 4 services, or both. For hospitalizations in both categories, we calculated total out-of-pocket Presence of cost-sharing waivers. The database did not report whether COVID-19 1 8 0 hospitalizations were covered by plans with cost-sharing waivers. However, as noted below, the 1 8 1 vast majority of hospitalizations in our sample did not have cost-sharing for facility services. While this might suggest that most hospitalizations were covered by insurers that waived cost- sharing for facility services -that is, that the absence of cost-sharing for facility services implied 1 8 4 the presence of a waiver -a potential alternative explanation is that most patients had already facility services, that would support the notion that cost-sharing waivers, rather than meeting out-1 9 3 of-pocket maximums, drove the low observed incidence of cost-sharing for facility services. We also explored whether it was reasonable to assume that hospitalizations with out-of-1 9 5 pocket spending for facility services were not covered by insurers with cost-sharing waivers for 1 9 6 these services (i.e., that the presence of cost-sharing for facility services implied the absence of a 1 9 7 waiver -the inverse of the assumption above). To evaluate this assumption, we compared the 1 9 8

incidence of out-of-pocket spending for facility services between COVID-19 hospitalizations and hospitalizations included also had a COVID-19 diagnosis code (U017).

Statistical analysis. We used descriptive statistics to assess patient characteristics, length 230 were excluded because the insurer was secondary, 63 because data on billing provider type 2 1 8 were missing, and 3 because out-of-pocket spending data were missing. Overall, 296 (6.8%) for Medicare Advantage patients, 1,432 (53.1%) were for females. Mean length of stay was 9.2 2 2 7 days (SD 8.9); 1,212 (44.9%) hospitalizations involved intensive care unit utilization.

Privately insured hospitalizations were most commonly covered by preferred provider 2 2 9 organization plans (47.0%). Mean and median allowed amounts for privately insured Out-of-pocket spending. Of the 1,377 and 2,698 hospitalizations for privately insured and services. Among these 63 and 36 hospitalizations, mean (SD) total out-of-pocket spending was 2 3 7 $3,840 (3,186) and $1,536 (1,402). In contrast, of the 1,377 and 2,698 hospitalizations for 2 3 8 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted May 30, 2021. ; 1 1 privately insured and Medicare Advantage patients, 981 (71.2%) and 1,324 (49.1%) had out-of-2 3 9 pocket spending for facility services, professional/ancillary services, or both. Among these 981 2 4 0 and 1,324 hospitalizations, mean total out-of-pocket spending was $788 (1,411) and $277 (363) 2 4 1 ( Table 2) . Of all 1,377 privately insured hospitalizations, 99 (7.2%) and 34 (2.5%) had total out-2 4 2 of-pocket exceeding $2,000 and $4,000. Of all 2,698 hospitalizations for Medicare Advantage 2 4 3 patients, the corresponding numbers were 7 (0.3%) and 5 (0.2%). Table 3 shows the incidence and magnitude of out-of-pocket spending for each of the 3 Consequently, insurer waivers could be heterogeneous, with some applying only to facility Insurers and clinicians might consider three steps to mitigate patient financial liability for 2 9 0 professional/ancillary services related to COVID-19 hospitalizations. First, insurers with no cost-2 9 1 sharing waiver or with waivers of limited scope could consider implementing a comprehensive 2 9 2 waiver, for example one that covers all services on or between the admission and discharge dates 2 9 3 of hospitalizations. Second, insurers that already have comprehensive waivers could work to 2 9 4 ensure appropriate implementation. Finally, clinicians could encourage patients to contest any 2 9 5 bills for professional/ancillary services that should be covered under an insurer's cost-sharing 2 9 6 waiver.

In this study, 4.6% and 1.5% of hospitalizations for privately insured and Medicare Advantage patients had out-of-pocket spending for facility services. Among these 2 9 9

hospitalizations, mean total out-of-pocket spending was $3,840 and $1,536, respectively. If the 3 0 0 presence of out-of-pocket spending for facility services implies the absence of an insurer cost- sharing waiver for these services -as suggested by the fact that the vast majority of influenza 3 0 2 hospitalizations had cost-sharing for facility services -our findings suggest that out-of-pocket 3 0 3 burden for COVID-19 hospitalizations could be large without insurer cost-sharing waivers. This Anthem and United Healthcare, allowed their cost-sharing waivers for COVID-19 3 0 6 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted May 30, 2021. ;

1 4 hospitalizations to expire. 9,10 Analyses suggest patients covered by these insurers could now face 3 0 7 substantial financial burden for COVID-19 hospitalizations, particularly the privately insured.

A strength of this study its use of national, fully-adjudicated claims data. Such data are 3 0 9

typically considered complete after a six-month time lag, meaning claims through the latter part This study also has limitations. First, despite strong indirect evidence, we cannot prove 3 1 5 that COVID-19 hospitalizations in this study were mostly covered by plans with cost-sharing 3 1 6 waivers. Second, if patients did not pay the amounts they were billed, the incidence of actual out-3 1 7 of-pocket spending would differ from the incidence estimated by this study. However, the 3 1 8 amount billed to patients still illustrates the financial burden patients may face without cost- responsibility for the integrity of the data and the accuracy of the data analysis. Funding source: Funding for purchasing IQVIA data was partially provided by the Susan B. . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) Foundation; 2020. 4 1 0 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted May 30, 2021. ; .

It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted May 30, 2021. Table 3 . Incidence and magnitude of out-of-pocket spending for professional/ancillary services among COVID-19 hospitalizations, IQVIA PharMetrics for Academics Database

Had OOP spending for facility services, professional/ancillary services

Medicare Advantage, COVID-19 (n = 2,698)

Had OOP spending for facility services 36 (1.3%) $1,536 (1,402) $1

Had OOP spending for facility services, professional/ancillary services

Had OOP spending for facility services 51 (83.6%) $

Had OOP spending for facility services, professional/ancillary services

Had OOP spending for facility services 159 (89.3%) $1,226 (708) $1,117 (665) $

Had OOP spending for facility services, professional/ancillary services, or both 173 (97.2%) $1,150 (728) $

OOP -out-of-pocket a See Appendix 2 for codes used to identify facility and professional/ancillary services. Facility services were those billed by hospitals for services such as accommodation. Professional/ancillary services were those billed by clinicians and ancillary providers

a Professional/ancillary services include those submitted by clinicians and those from ancillary service providers, such as ambulance providers. See Appendix 2 for details b Services from miscellaneous providers, such as a durable medical equipment provider or dialysis center c Includes services submitted by clinicians with a hospital place of service but no procedure code for evaluation and management or diagnostic services

11. The Heroes Act, H.R. 6800 (2020). https://www.kff.org/medicare/issue-brief/a-dozen-facts-about-medicare-advantage-in-3 9 82019. Accessed July 8, 2020.

Privately insured (n = 1,377 hospitalizations) Medicare Advantage (n = 2,698 hospitalizations)