LTC: Can More Home and
Community Based Support Save
Community-Based
Money
y While Improving
p
g Care?
Dual Eligibles in Maryland
Annual Research Meeting
Charles Milligan
June 28
28, 2010
Preview of Presentation

Overview of Project

Results

Selected Policy Implications
Overview of Project
The overall project goals

Look at cross-payer effects for dual eligibles in
Maryland

One subgroup analysis: Look at the cross-payer effects
for dual eligibles who meet nursing facility level of care
(NF LOC), both in the community and in institutions

The major non-Developmental Disabilities HCBS waiver
in Maryland is the Older Adults Waiver (OAW)

The OAW waiver recipients were compared to matched
individuals in the community and in LT-NFs
Propensity score matching
was done for the comparisons

The “treatment” was receipt of HCBS waiver services in the
potential treatment g
group:
p 12-month continuously
y
OAW. The p
enrolled dual eligibles aged 50+ who were enrolled in the OAW
the entire year (2006) (n=1,759)

Potential controls (community): 12-month continuously enrolled
dual eligibles aged 50+ who did not receive Medicaid long-term
supports and services in 2006 (n=19,095)

Potential
o e a co
controls
o s ((LT-NF):
) 12-month
o
co
continuously
uous y e
enrolled
o ed dua
dual
eligibles aged 50+ who had ≥ 30 days of Medicaid-paid NF care
just prior to January 1, 2006, and had NF care ≥ 10 months of
2006 (n=6,336)
(n 6 336)
-5-
Propensity score matching
covariates utilized

Covariates included in propensity score
estimation
ti ti









Age
Gender
Race
CMS-HCC relative value
20 Chronic Condition Warehouse condition indicators
Disability as reason for original Medicare entitlement indicator
Frailty indicator (diagnostic-based,
diagnostic based Hopkins ACG system)
ESRD indicator
Months of full Medicaid coverage
-6-
Outcome of propensity score
matching

Of the potential treatment group in the OAW (n=1,759):

1,440
1
440 matches were found among the potential
community control group population of 19,095, so the
sub-group
g p analysis
y
p
proceeded with the 1,440 OAW
recipients and the 1,440 individuals in the community
control group with whom they “matched”

1,731 matches were found among the potential LT-NF
control group population of 6
6,336,
336 so the sub-group
sub group
analysis proceeded with the 1,731 OAW recipients and
the 1,731 in the LT-NF control g
group with whom they
y
“matched”
-7-
Results
Medicare payments were nearly identical
f OAW recipients
for
i i
and
d the
h community
i
control group . . .
MEDICARE Benefit Payments, PMPM, by Service
OAW (Treatment)
Community (Control)
$
$5,000
Although the overall costs were similar, the
OAW g
group
p had fewer hospital
p
readmissions,,
fewer/shorter SNF stays, more home health,
more DME, and fewer ER encounters, which
suggests the OAW recipients received better
coordination with Medicare.
Cla
aim Payments
$4,000
$3,000
$2 000
$2,000
$1,219 $1,302
$1,000
$548
$586
$425
$60
$90
$79
$44
$32
$451
$93
$75
$38
$0
Total
Hospital
SNF
Home Health
Hospice
Physician/
Outpatient
DME
Source: Tucker, A., & Johnson, K. (2010). Cross-Payer Effects on Medicare Resource Use: Lessons for Medicaid Administrators. Baltimore, MD: The Hilltop
Institute.
Notes: Maryland OAW (treatment) and community (control) samples of 1,440 full-benefit duals aged 50 and older, enrolled for 12 months (with no group health
coverage) in 2006.
-9-
. . . while Medicaid payments were far
hi h for
higher
f the
th OAW recipients
i i t th
than th
the
community control group . . .
MEDICAID Benefit Payments, PMPM, by Service
OAW (Treatment)
Community (Control)
$5 000
$5,000
Cla
aim Payments
$4,000
$3,000
$2,784
$2,626
$2 000
$2,000
$1,000
$360
$0
$0
Total
$18
Hospital
$0
$151
$11
Nursing Facility
$0
$92
$148
$16
$9
$72
Community
Hospice
Physician/
DME
Supports & Services
Outpatient
Source: Tucker, A., & Johnson, K. (2010). Cross-Payer Effects on Medicare Resource Use: Lessons for Medicaid Administrators. Baltimore, MD: The Hilltop
Institute.
Notes: Maryland OAW (treatment) and community (control) samples of 1,440 full-benefit duals aged 50 and older, enrolled for 12 months (with no group health
coverage) in 2006. Medicare crossover payments paid by Medicaid not included.
-10-
. . . and as a result, the OAW recipients
were far
f more expensive
i
than
th
th
the
community control group, in total dollars.
MEDICARE and MEDICAID Benefit Payments,
PMPM, by Service
OAW (Treatment)
(T t
t)
C
Community
it (Control)
(C t l)
$6,000
Claim
m Payments
$5,000
$4,000
$4,003
$3,000
$2,000
$2,705
$1,662
$1,000
$548
$604
$573
$
$60
$241
$
$55
$0
Total
Hospital
SNF/NF
Community (incl.
Home Health)
$32
$467
$185
Hospice
$84
Physician/
Outpatient
$110
DME
Source: Tucker, A., & Johnson, K. (2010). Cross-Payer Effects on Medicare Resource Use: Lessons for Medicaid Administrators. Baltimore, MD: The Hilltop
Institute.
Notes: Maryland OAW (treatment) and community (control) samples of 1,440 full-benefit duals aged 50 and older, enrolled for 12 months (with no group health
coverage) in 2006. Medicare crossover payments paid by Medicaid not included.
-11-
Medicare payments were $441 higher
PMPM ffor th
the OAW group th
than th
the LT
LTNF control group . . .
MEDICARE Benefit Payments, PMPM, by Service
OAW (Treatment)
LT-NF (Control)
$5 000
$5,000
Claim Payments
$4,000
$3,000
$2,000
$
,
$1,227
$1,000
$786
$545
$403
$312
$68
$89
$76
$0
Total
Hospital
SNF
$0
Home Health
$
$44
$354
$78
$2
Hospice
Physician/
Outpatient
$41
DME
Source: Tucker, A., & Johnson, K. (2010). Cross-Payer Effects on Medicare Resource Use: Lessons for Medicaid Administrators. Baltimore, MD: The Hilltop
Institute.
Notes: Maryland OAW (treatment) and community (control) samples of 1,731 full-benefit duals aged 50 and older, enrolled for 12 months (with no group health
coverage) in 2006.
-12-
. . . while Medicaid payments were
$2 055 PMPM hi
$2,055
higher
h for
f th
the LT
LT-NF
NF
group, compared to the OAW group . . .
MEDICAID Benefit Payments, PMPM, by Service
OAW (Treatment)
$4 835
$4,835
$5 000
$5,000
LT-NF (Control)
$4 832
$4,832
Claim Payments
$4,000
$3,000
$2,780
$2,621
$2,000
$
,
$1,000
$0
$0
Total
$0
Hospital
$1
$0
Nursing Facility
Community
Supports & Services
$0
$2
Hospice
$151
$1
Physician/
Outpatient
$8
$0
DME
Source: Tucker, A., & Johnson, K. (2010). Cross-Payer Effects on Medicare Resource Use: Lessons for Medicaid Administrators. Baltimore, MD: The Hilltop
Institute.
Notes: Maryland OAW (treatment) and community (control) samples of 1,731 full-benefit duals aged 50 and older, enrolled for 12 months (with no group health
coverage) in 2006. Medicare crossover payments paid by Medicaid not included.
-13-
. . . and in total dollars, the OAW was
far less expensive than the LT-NF.
MEDICARE and MEDICAID Benefit Payments,
PMPM, by Service
OAW (Treatment)
(T t
t)
$6,000
$5,621
$4,908
Claim
m Payments
$5,000
$4,000
LT NF (Control)
LT-NF
(C t l)
$4,007
$3,000
$2,710
$2,000
$1,000
$554
$545
$312
$
$69
$0
$0
Total
Hospital
SNF
Home Health
$44
$355
$4
Hospice
Physician/
Outpatient
$86
$41
DME
Source: Tucker, A., & Johnson, K. (2010). Cross-Payer Effects on Medicare Resource Use: Lessons for Medicaid Administrators. Baltimore, MD: The Hilltop
Institute.
Notes: Maryland OAW (treatment) and community (control) samples of 1,731 full-benefit duals aged 50 and older, enrolled for 12 months (with no group health
coverage) in 2006. Medicare crossover payments paid by Medicaid not included..
-14-
In sum, in total dollars, the OAW group is far more
expensive than the community control group and
far less expensive than the LT-NF control group.
MEDICARE and MEDICAID Benefit Payments,
PMPM, by Source of Controls
OAW (Treatment)
Controls
$6,000
$5,621
Claim Payments
$5,000
$4,000
$4,007
$4,003
$3,000
$2,000
$1,662
$1,000
$0
Community controls
LT-NF controls
Source: Tucker, A., & Johnson, K. (2010). Cross-Payer Effects on Medicare Resource Use: Lessons for Medicaid Administrators. Baltimore, MD: The Hilltop
Institute.
Notes: Maryland OAW (treatment) and community (control) samples of 1,731 full-benefit duals aged 50 and older, enrolled for 12 months (with no group health
coverage) in 2006. Medicare crossover payments paid by Medicaid not included.
-15-
While the Medicare payment difference is
negligible
li ibl in
i th
the community,
it th
the Medicare
M di
program saved $$ when people were in NFs.
MEDICARE and MEDICAID Benefit Payments, PMPM,
by Source of Controls
OAW (Treatment)
(T t
t)
C t l
Controls
$4,835
$5,000
Claim Payments
$4,000
$2,784
$3,000
$2,780
$2,000
$1,219
$1,302
$1,227
$786
$1,000
$360
$0
Community
controls:
Medicare
Community
controls:
Medicaid
LT-NF
controls:
Medicare
LT-NF
controls:
Medicaid
Source: Tucker, A., & Johnson, K. (2010). Cross-Payer Effects on Medicare Resource Use: Lessons for Medicaid Administrators. Baltimore, MD: The Hilltop
Institute.
Notes: Both sets of samples: full-benefit duals aged 50 and older, enrolled for 12 months (with no group health coverage) in 2006. Maryland OAW and Community
samples: n=1,440; Maryland OAW LT-NF samples: 1,731. Medicare crossover payments paid by Medicaid not included.
-16-
Keep in mind: most discrete SNF/NF
stays
t
begin
b i as a post-acute
t
t M
Medicare
di
SNF stay . . .
Source of Payment at Initial Admission
Admission, Discrete Stays
Private/Other
11%
Medicaid
15%
Medicare
74%
Hilltop refined MDS data
data, Discrete Stays in Maryland
Maryland, 1999
1999-2008
2008
-17-
. . . and the vast majority of
“
“extended”
d d” SNF/NF stays begin
b i as a
post-acute Medicare SNF stay.
Source of Payment at Initial Admission, Extended Stays
Medicaid
6%
Private/Other
11%
Medicare
83%
Hilltop refined MDS data, Extended Stays in Maryland, 1999-2008
Selected Policy Implications
No. 1: Medicare and Medicaid
financing don’t align to promote
HCBS

Medicare spent less when a dual eligible was in an NF ($786 PMPM)
than it did when a comparable person was in the OAW in the community
($1,227 PMPM) (and Medicare spent about the same for a comparable
person in the community who was not in the OAW or any waiver)

Medicaid spent less when a dual eligible was in the OAW waiver ($2,780
PMPM) than it did when a comparable person was in an NF ($4,835
PMPM))

Overall, placement in the OAW costs substantially less ($4,007 PMPM)
than p
placement in an LT-NF (($5,621 PMPM))

To align the financial incentives toward community placement,
integrated care for dual eligibles must involve one entity that bears the
Medicare and Medicaid risk
No. 2: Because the vast majority of LT-NF
g with a Medicare stay,
y,
admissions begin
community integration for dual eligibles must
engage Medicare

From 1999-2008, 74% of all discrete admissions in MD
to an SNF/NF began as a Medicare stay
stay, whereas only
15% began as a Medicaid stay (total stays = 648,774)

Over the same period, 83% of all “extended” stays in
MD began as a Medicare stay (total stays = 384,156)

This emphasizes the need to engage Medicare to
promote community-based care of LT-NF admissions
for dual eligibles
No. 3: The OAW only is cost
effective for Medicaid when it truly
avoids an LT-NF placement

Offering an OAW waiver slot ($2,780 PMPM) to an
individual who is certain to become an LT
LT-NF
NF resident
($4,835 PMPM) is cost-effective for Medicaid

Offering an HCBS waiver slot to an individual ($2,784
PMPM) who otherwise would remain in the community
anyway ($360 PMPM) is more expensive to Medicaid

This emphasizes the need for good assessment
processes to award slots to people at-risk for LT-NF
Contact Information
Charles Milligan
Executive Director
The Hilltop Institute
University of Maryland
Maryland, Baltimore County (UMBC)
410.455.6274
cmilligan@hilltop.umbc.edu
www.hilltopinstitute.org
-23-