Transcript PowerPoint Slides

Dollars to Doughnuts:
Predicting Prescription Drug
Costs of Beneficiaries and the
Medicare Program Under Part D
M. Christopher Roebuck, Caremark
Dominick Esposito, Mathematica Policy
Research, Inc.
Meredith Lewis, Caremark
Jan Berger, Caremark
Academy Health Annual Research Meeting,
Seattle, June 26, 2006
This presentation contains confidential and proprietary information of Caremark and
cannot be reproduced, distributed, or printed without written permission from Caremark.
©2006 Caremark. All rights reserved.
Research Objectives
 Examine drug utilization and out-of-pocket (OOP)
costs of Medicare beneficiaries using a Medicare
prescription drug discount card that includes
beneficiaries who qualified for the Transitional
Assistance Program (TAP)
 Simulate Medicare beneficiary OOP costs
(excluding premiums) and costs to Medicare from
the standard Medicare Part D benefit
 Understand the factors associated with these
simulated costs and the probability of being in the
doughnut hole (coverage gap)
2
Caremark proprietary and confidential information. Not for distribution.
Medicare Drug Discount Card
 Created via the Medicare Modernization Act
 Enrolled nondual eligible beneficiaries into the
temporary program June 2004 to December 2005
 Provided a $600 annual subsidy to seniors with
income below 135% of the federal poverty level
(TAP)
 May have charged cardholders an enrollment fee of
up to $30 per calendar year
3
Caremark proprietary and confidential information. Not for distribution.
Data
 Consist of eligibility and prescription claims data for
enrollees with one of 34 separate Medicare drug
discount cards managed by Caremark
 Comprised of beneficiaries enrolled for six or more
months with one or more pharmacy claims between
June 2004 and November 2005 (n=37,425)
 Generated Pharmacy Health Dimensions (PHD), a
pharmacy-based risk index that categorizes
prescription data into 62 disease indicators1
Source: 1 Powers, C.M., Meyer, C.M., Roebuck, M.C. and B. Vaziri. 2005. “Predictive Modeling of Total Healthcare Costs Using Pharmacy Claims
Data: A Comparison of Alternative Econometric Cost Modeling Techniques.” Medical Care 43(11): 1065-1072.
Caremark proprietary and confidential information. Not for distribution.
4
Sample Characteristics
TAP
(n=17,317)
Non-TAP
(n=20,108)
Total
(n=37,425)
26%
38%
33%
Mean age
76
75
76
Number of conditions
2.5
2.0
2.2
Hypertension
60%
45%
52%
High cholesterol
27%
30%
27%
Diabetes
19%
14%
16%
Hypothyroidism
17%
13%
15%
Osteoarthritis
18%
12%
15%
Allergies
14%
10%
12%
Depression
13%
10%
11%
Male
Note: All differences in means across TAP and Non-TAP groups are statistically significant (p<0.01) using Kruskal-Wallis equality of populations
test, except high cholesterol.
Caremark proprietary and confidential information. Not for distribution.
5
Annualized Utilization and Costs Under
Medicare Drug Discount Card Program
TAP
(n=17,317)
Non-TAP
(n=20,108)
Total
(n=37,425)
Mean annual number of
prescriptions
Brand
11
8
9
Generic
12
7
10
Total
23
15
19
Enrollee paid (OOP)
$256
$781
$538
Medicare paid (TAP)
$672
$0
$311
Total
$928
$781
$849
Mean annual drug costs
Note: All differences in means across TAP and Non-TAP groups are statistically significant (p<0.01) using Kruskal-Wallis equality of populations test.
Caremark proprietary and confidential information. Not for distribution.
6
Standard Medicare Part D Benefit
Enrollee OOP Cost Share
$0 $250
$2250
$5100
100% co-insurance
25% co-insurance
$2/$5 co-pay or 5% co-insurance
7
Caremark proprietary and confidential information. Not for distribution.
Standard Medicare Part D Benefit
Simulation
 Applied Standard Medicare Part D benefit costshare structure to annualized total drug costs (plus
a 3.5% price inflation adjustment from 2005 to
2006)
 Expected utilization increases under insurance
 Applied variable “induction factors” of between
0.701 and 1.252 for prescription drugs
 Estimated each $1 decrease in OOP costs to induce
between $0.70 and $1.25 of increased drug spending
Sources: 1. Mays, J., Brenner, M., Neuman, T., Cubanski, J., and G. Claxton. November, 2004. “Estimates of Medicare Beneficiaries’ Out-of-Pocket Drug
Spending in 2006: Modeling the Impact of the MMA.” Henry J. Kaiser Family Foundation. Menlo Park, CA.
2. American Academy of Actuaries. May 1995. “Medical Savings Accounts: Cost Implications and Design Issues.” Washington, DC.
8
Caremark proprietary and confidential information. Not for distribution.
Beneficiary OOP Costs
Actual Drug Discount Card
Simulated Part D (no induction)
Simulated Part D (induction factor: 0.70)
Simulated Part D (induction factor: 1.25)
Simulated Part D (induction factor: 1.00)
$900
$771
$800
$781
$725
$690
$700
$685
$600
$555
$500
$441
$655
$625
$601
$410
$576
$538
$424
$400
$300
$256
$200
$100
$0
TAP
Non-TAP
Total
9
Caremark proprietary and confidential information. Not for distribution.
Percent of Beneficiaries in the
Doughnut Hole
Simulated Part D (no induction)
Simulated Part D (induction factor: 0.70)
Simulated Part D (induction factor: 1.00)
Simulated Part D (induction factor: 1.25)
30.0%
26.7%
25.0%
23.4%
21.9%
20.6%
20.0%
20.0%
18.5%
16.0%
15.7%
15.5%
15.0%
10.0%
8.1%
6.7%
5.1%
5.0%
0.0%
TAP
Non-TAP
Total
10
Caremark proprietary and confidential information. Not for distribution.
Multivariate Analyses
 Estimated generalized linear models (GLM;
gamma/log link) for:
 Simulated OOP costs under Medicare Part D
 Simulated Medicare payments under Medicare Part D
 Estimated a probit model for the likelihood of being in
the doughnut hole
 Explanatory variables included:
- Age, gender, geographic region
- TAP status
- Generic dispensing rate
- 62 PHD disease indicators
11
Caremark proprietary and confidential information. Not for distribution.
Multivariate Results
 TAP (low income) status is associated with:
 $58 to $86 higher beneficiary costs
 $199 to $238 higher Medicare payments
 Almost no change in the probability of being in the
doughnut hole
 A 10-percentage point increase in the generic
dispensing rate is associated with a:
 $41 to $55 decrease in beneficiary costs
 $62 to $71 decrease in Medicare payments
 1.8 to 2.6 percentage point reduction in the
probability of being in the doughnut hole
12
Caremark proprietary and confidential information. Not for distribution.
Conclusions/Policy Implications
 In choosing whether or not to enroll in Medicare Part D,
beneficiaries will compare annual premiums with the
expected payout of the Medicare program
 Risk-neutral beneficiaries will enroll at monthly premiums
below the $52 to $60 range
 Faced with higher OOP costs, low-income beneficiaries
without low-income subsidies may reduce their drug
utilization, potentially resulting in adverse health effects
(e.g., those not passing the asset test)
 To reduce costs, Medicare should actively promote
generic substitution
 With 33 million enrollees, the Medicare program could save
more than $2.0 billion to $2.3 billion annually by increasing
the generic dispensing rate by 10%
13
Caremark proprietary and confidential information. Not for distribution.
Strengths and Limitations
 Strengths:
 Actual claims experience, not self-reported use
 Actual OOP amounts
 Data on a low-income population (TAP)
 Limitations:
 Enrollees with zero claims excluded from the analysis
 Discount card may not always have been used
 Simulating the standard Medicare Part D benefit,
although most have selected nonstandard plans
14
Caremark proprietary and confidential information. Not for distribution.
Thank you.
Contact:
M. Christopher Roebuck
(410) 785-2136
[email protected]
15
Caremark proprietary and confidential information. Not for distribution.