Hospice Carve-out Uneven Prevalence of Medicare Advantage Enrollment - - PowerPoint PPT Presentation

www.rti.org

RTI International is a registered trademark and a trade name of Research Triangle Institute.

Unintended Consequences of the Hospice Carve-out

Uneven Prevalence of Medicare Advantage Enrollment across Hospices

Ila H. Broyles, PhD RTI International

Qinghua Li, PhD Nan Tracy Zheng, PhD Samantha Zepeda, BS Natalie Chong, BA Ann Larsen, MS Alon Evron, MA Micah Segelman, PhD Jennifer Frank, MPH Franziska Rokoske, PT, MS RTI International Cindy Massuda Contracting Officer Representative Centers for Medicare and Medicaid Services

Acknowledgements

Co-Authors Qinghua Li, PhD Nan Tracy Zheng, PhD Samantha Zepeda, BS Natalie Chong, BA Ann Larsen, MS Alon Evron, MA Micah Segelman, PhD Jennifer Frank, MPH Franziska Rokoske, PT, MS This project was funded in part by the Centers for Medicare & Medicaid Services under contract no. HHSM-500-2013-13015I. The statements contained in this presentation are solely those of the authors and do not necessarily reflect the views or policies of the Centers for Medicare & Medicaid Services. RTI assumes responsibility for the accuracy and completeness of the information contained in this report.

Background: Medicare Advantage and the Hospice Carve-out

Overview of Medicare Advantage and Carve-out Policy

• Medicare Advantage (MA) is a voluntary program that allows

beneficiaries to receive Medicare benefits through a private health insurer.

• Private insurers receive a capitated payment to cover *almost* all

Medicare benefits for enrollees.

• Unlike other parts of the MA benefit, MA enrollees who elect hospice

receive hospice benefits through fee-for-service (FFS) Medicare, so their hospice benefits are “carved-out” of their MA benefits.

• Recommendations about ending the carve-out (MedPAC, etc.)

highlight some of the trade-offs in the current model.

Current Rules: Medicare Advantage and Hospice

• Beneficiaries who receive MA coverage for their Medicare benefits but

wish to elect hospice are not required to drop their MA coverage:

– For all services related to their terminal illness, Medicare pays for their

hospice benefit through the per diem paid to the hospice provider.

– For services not related to their terminal illness (including those of the

managed care attending physician), Medicare pays fee-for-service (FFS).

– Finally, for any supplemental benefits (which are not original Medicare

benefits) and Part D (if-MA-PD plan), the MA plan is responsible for continued coverage and payment and Medicare pays the MA plan a reduced capitated payment.

• Some beneficiaries may retain their MA coverage, elect hospice, but

subsequently decide to revoke their election. In that case, Medicare pays FFS for services received after the hospice revocation date until the first day of the following month.

Implications of the Carve-out Policy

• MA enrollment has dramatically risen in the last decade from

13% in 2003 to 31% in 2015 of Medicare beneficiaries.

• This prevalence is uneven across states, ranging from <10% to

>50% of Medicare beneficiaries in each state receiving their Medicare benefits through MA plans.

Research Objectives

• To understand the geographic variation in hospice patients’

enrollment in MA before, during, and after their hospice election.

• To understand whether certain kinds of hospices are more likely to

serve MA enrollees.

Current Picture: Medicare Advantage and the Hospice Carve-out

Sample and Study Design

• Data Sources:

– Medicare hospice claims – POS File with hospice characteristics merged onto patient-stays – Enrollment Database with MA enrollment and other beneficiary

characteristics

• Sample: 3.5 million hospice enrollees with hospice discharges in

2013-2015 and who elected hospice after January 1, 2012

• Methods:

– Descriptive analysis: patient-level MA enrollment before, during, and after

hospice election

– Multivariate analysis: hospice-level analysis of proportion of patient-stays

with MA enrollment

Prevalence of MA Enrollment among Hospice Enrollees

• Roughly one-third of hospice

enrollees had MA enrollment concurrent with their hospice enrollment

– Almost all of these patients were

enrolled in MA for their entire hospice election

• Most hospice enrollees with

concurrent MA enrollment were previously enrolled in MA

69% 29% 2% <1%

MA Enrollment among Hospice Enrollees

No MA Enrollment Continuous MA Only MA Before Only MA Concurrent

Geographic Variation in MA Enrollment: Hospice Enrollees

Large state variation in hospice enrollees with MA enrollment, ranging from >50% of hospice patients in Hawaii, Minnesota, and Puerto Rico to <10% in Delaware, New Hampshire, Vermont, and Wyoming STATE HEAT MAP INSERTED HERE

Patient Characteristics by MA Enrollment

Hospice Patient Characteristics MA Enrollees (N= 1,036,496) Non-MA Enrollees (N= 2,510,929) White* 84% 88% Female* 56% 58% Received care from… Non-profit Hospice* 42% 41% For-profit Hospice 45% 45% Urban Hospice* 92% 86% Hospice with Participation Pre-2000* 34% 33%

MA enrollees listed here have some MA enrollment during the time they elected hospice (“concurrent MA”) *Chi-squared test indicated significant differences, p<.01

Hospice Characteristics and MA Enrollment

• Analytic Approach: Regression Analysis

– Sample: 4,395 hospices with patients discharged in 2013-2015 – Dependent Variable: Proportion of the hospice’s patients with MA

enrollment

– Independent Variables:

• Average age of patients
• Proportion of female patients
• Proportion of non-hispanic white patients
• Urban (categorical)
• Non-profit/for-profit/government owned (categorical)
• Participation prior to 2000 (categorical)
• Hospice size (categorical by quintile)

– Linear regression with robust standard errors, state fixed-effects

Hospice Characteristics and MA Enrollment

• Wide variation in hospices’ proportion of patients with MA coverage

(0-100%) with a median of 25% of patients

– 53 hospices with >80% of patients with MA coverage had a median total

stays in the three year period of 100, geographically diverse, etc.

Hospice Characteristics and MA Enrollment

Independent Variables Coefficient Proportion Non-Hispanic White

• 0.138**

Proportion Female

• 0.0122

Average Age

• 0.00096

Non-profit Hospice 0.0081 Government Hospice 0.0057 Urban Hospice 0.068** Hospice Participation Year Pre-2000 0.016**

*Indicated significant differences, p<.05

Adjusting for the state of hospice location, increased MA penetration was associated with certain hospice characteristics:

• Urban providers had a 6.8 percentage point increase in the percent of

patients with MA (p<.05).

• Hospices operating for >15 years had 1.6 percentage point increase in the

percent of patients with MA (p<.05).

• The largest 20% of hospices had a 5.4 percentage point increase in the

percent of patients with MA relative to the smallest 20% of hospices (p<.05)

MA Enrollment after Hospice Live Discharge

• MA after live discharge

– 30% of patients live discharged were subsequently enrolled in MA

• 5% of all patient-stays

– 99% of these had MA coverage during their hospice election and most for

their entire hospice election

– Hospices’ percent of total patient-stays with MA after live discharge:

• Mean: 7.5%
• Median: 4.6%
• Max: 100%

Implications: Medicare Advantage and the Hospice Carve-out

MA Enrollment and Hospice Delivery

• Observed variability in MA penetration across hospices may affect

hospice delivery

– MA enrollees may have access to a different set of palliative care and

supportive services compared to FFS patients before hospice election

– Payments to MA plans for MA enrollees who elect hospice may alter

hospice election patterns

• MA enrollment may then affect the timing of hospice enrollment and

the types of patients who elect hospice

– Analysis suggests that MA patients are more likely to enroll with certain

kinds of hospices, although the reason for that is not clear

– Unknown whether MA penetration by hospice also varies with quality of

hospice care

• Public reporting of hospice quality information in the HQRP may amplify those

patterns, if they exist

MA Enrollment and Hospice Delivery

• Variability in MA penetration across hospices may affect future

research and quality measure development

– Utilization and diagnosis information is commonly used for risk adjustment – Data suppression for MA enrollees before and after hospice election due to

lack of FFS claims

– Limited utilization and diagnosis information (due to suppression) may

affect future risk adjustment

More Information

Ila Broyles 919-485-2759 ibroyles@rti.org Franziska Rokoske 919-541-8833 frokoske@rti.org