Health Savings Accounts: Early estimations on national take-up from 2003 MMA and future policy proposals Stephen T Parente Roger Feldman Jean Abraham Jon B Christianson Funded by the Robert Wood Johnson Foundation Health Care Financing and Organization Initiative (HCFO) and the Department of Health and Human Services
38
Embed
Health Savings Accounts: Early estimations on national take-up from 2003 MMA and future policy proposals Stephen T Parente Roger Feldman Jean Abraham Jon.
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Health Savings Accounts: Early estimations on national take-up from 2003 MMA and future policy proposalsStephen T Parente
Roger FeldmanJean AbrahamJon B Christianson
Funded by the Robert Wood Johnson Foundation Health Care Financing and Organization Initiative (HCFO) and the Department of Health and Human Services
Presentation Overview Consumer Driven Health Plan
Overview Research Questions Data & Analytic Approach Plan Choice Estimation Results MEPS Data Policy Simulation Results Implications Next Steps
‘Classic’ CDHP Model – Definity Health
Definity Definity HealthHealthCareCare
AdvantageAdvantage
Web- and Web- and Phone-Phone-Based Based ToolsTools
Health ToolsHealth Toolsand Resourcesand Resources
Health Tools and Resources• Care management
program• Internet enabled
Health Coverage• Preventive care covered
100%• Annual deductible• Expenses beyond the
PCA
Personal Care Account (PCA)• Employer allocates PCA1
• Member directs PCA• Roll over at year-end • Apply toward deductible2
Annual Annual DeductibleDeductible
Annual Annual DeductibleDeductible
Pre
ven
tive
Care
10
0%
Pre
ven
tive
Care
10
0%
Health Health CoverageCoverage
An
nu
al
Ded
uct
ible
1 Employer selects which expense apply toward the Health Coverage annual deductible.2 Paid out of employer’s general assets.
PCAPCAPCAPCA
$$
The HSA ModelAn HSA is a special account owned by the individual where tax-free contributions to the account are used to pay for current and future medical expenses.
Used with High Deductible Health Plans (HDHPs)
Bush Administration has proposed refundable tax credits for individuals to purchase plans with HSAs.
Qualified HDHPs now offered by: UnitedHealth, the Blues, Aetna (w/preventive meds), Cigna, Humana and Kaiser Permanente
Annual Annual DeductibleDeductible
Annual Annual DeductibleDeductible
Pre
ven
tive C
are
P
reven
tive C
are
1
00
%1
00
%
Health Health CoverageCoverage
An
nu
al
Ded
uct
ible
HSAHSAHSAHSA
$$
Nearly National Appeal: States where the study employers’ 1st year CDHP take-up was >5%
>5%
0.1 - 5%
0%
Take-up
Data based on 6 large employers representing ~250,000 covered lives with initial year offerings from 2001 to 2oo3.
Economic Analysis Motivation ‘Donut hole’ and savings account are new
‘prices’ to consider in the demand for health insurance with expected negative and positive responses, respectively.
Price sensitivity to different benefit options (i.e., premium, account, donut, coinsurance) could significantly affect take-up of CDHPs.
Builds on existing literature of high deductible health plans (HDHPs) (Keeler, Newhouse, Phelps, 1977). HSAs introduce a new kink in the budget constraint.
Conceptual Model of CDHP
A to B: Care AccountB to C: DeductibleC to D: Catastrophic insurance w/no coinsurance
Money
Medical CareSpending
HRA/HSA deductible
d
a b
cCDHP Budget
Research Questions How does the introduction of Consumer
Directed Health Plans (CDHPs) into mainstream health insurance affect plan choice?
What is the expected take-up rate of HSAs in the individual market from the 2003 MMA?
What is the impact of the Administration’s proposed HSA subsidies? Take-up rate of HSAs Impact on the uninsured Cost of the subsidy
What is the impact of other possible subsidy designs?
Data Sources 2002 health plan choice data from 3 large employers
participating in a Robert Wood Johnson Foundation funded study on CDHPs Employee premium, deductible, coinsurance, worker’s
Household Component: All adults age 19-64 not enrolled in public insurance programs and not full-time students during Round 1 Demographic, employment, and health insurance
information Linked Insurance Component: Subset of workers offered
employer coverage and their plan choices Plan type, premiums, contributions, coinsurance,
copayments and deductibles eHealthinsurance.com
Individual HSA plan information
Estimate plan offerings using linked data
Merge employer data
Estimate hedonic premium regression
Assign plan choices to full MEPS sample
Estimate plan choice regression
Use parameter estimates to predict plan choice probabilities for MEPSRe-scale take-up rates
* We use a Low HRA in our simulations for approximating an HSA benefit design.
Plan Choice – Conditional Logit (1 of 2)
Adjusted r-square: ~0.363, Reference PPO_High
StandardVariable Coefficient Error T-Statistic P-valueTax adjusted Employee Premium in $1,000 -1.7299 0.0665 -25.9996 <.0001Employee's Health Account in $1,000 0.6179 0.0974 6.3407 <.0001∆ Between Deductible and Health Account in $1,000 -0.8502 0.0272 -31.2867 <.0001Coinsurance (e.g., 15% = .15) -7.5675 0.5300 -14.2794 <.0001PPO Medium Plan Intercept=1, else=0 0.1153 0.0860 1.3404 0.1801PPO Low Plan Intercept=1, else=0 -1.0210 0.1311 -7.7887 <.0001High HRA Plan Intercept=1, else=0 -1.9816 0.1646 -12.0383 <.0001Low HRA* Plan Intercept=1, else=0 -1.7031 0.1252 -13.6042 <.0001HMO Plan Intercept=1, else=0 1.9338 0.1048 18.4506 <.0001Premium & Family Contract (0/1) Interaction 1.1067 0.0721 15.3418 <.0001PPO Medium & Age (in 100 years) Interacttion -1.5547 0.1673 -9.2926 <.0001PPO Low & Age (in 100 years) Interaction 0.5609 0.2917 1.9228 0.0545HRA & Age (in 100 years) Interaction -2.0759 0.3506 -5.9217 <.0001HSA_S/E & Age (in 100 years) Interaction -2.5626 0.2410 -10.6318 <.0001HMO & Age (in 100 years) Interaction -3.9740 0.1946 -20.4199 <.0001
N=28,737* We used Low HRA to create HSA 'predicting' coefficients for an individual version where theemployee pays all (S) and the employer offered version where the premium and the accountis heavily subsidized (E).
Plan Choice – Conditional Logit (2 of 2)
Adjusted r-square: ~0.363, Reference PPO_High
StandardVariable Coefficient Error T-Statistic P-valuePPO Medium & Income (in $1,000) Interaction -0.0005 0.0006 -0.8201 0.4122PPO Low & Income (in $1,000) Interaction 0.0038 0.0014 2.6689 0.0076HRA & Income (in $1,000) Interaction 0.0100 0.0009 11.6295 <.0001HSA_S/E & Income (in $1,000) Interaction 0.0102 0.0006 17.5883 <.0001HMO & Income (in $1,000) Interaction -0.0017 0.0007 -2.3806 0.0173PPO Medium & Female (0/1) Interaction 0.0852 0.0369 2.3100 0.0209PPOLow & & Female (0/1) Interaction -0.1566 0.0617 -2.5373 0.0112HRA & Female (0/1) Interaction -0.1665 0.0776 -2.1456 0.0319HSA_S/E & Female (0/1) Interaction -0.0278 0.0514 -0.5399 0.5893HMO & Female (0/1) Interaction -0.1892 0.0435 -4.3506 <.0001PPO Medium & Family Contract (0/1) Interaction 0.0218 0.0690 0.3154 0.7525PPO Low & Family Contract (0/1) Interaction 0.3809 0.0773 4.9275 <.0001HRA & Family Contract (0/1) Interaction -0.2979 0.1140 -2.6132 0.0090HSA_S/E & Family Contract (0/1) Interaction 0.0067 0.1144 0.0586 0.9533HMO & Family Contract (0/1) Interaction -0.3651 0.0909 -4.0175 <.0001
N=28,737
MEPS Data
We used the MEPS Household Component to identify adults 19-64 who were not full-time students and not enrolled in public insurance. We then subdivided them into 2 groups: (1) non-offered adult population and (2) workers with an offer of employer group coverage.
Descriptive Statistics(Means reported with SD in parentheses)
Non-Offered Population(N=36,816,103)
Offered Population(N=84,719,584)
Age 37.96(12.6)
40.54(10.84)
Female .477(.49)
.46(.49)
Income $17,974(21,286)
$40,454(28,741)
Single .56(.49)
.39(.49)
Assigning health plans to MEPS workers with an offer of employer coverage
1. Used the MEPS linked sample (n=3,127) to estimate a model for the number of health plans (e.g., PPO, HMO) offered to eligible workers.
2. Predicted the number and types of plans available to each offered worker in the MEPS Household Component using these model estimates (max of 4 plans).
3. Assigned plan attributes based on sample distributions from MEPS linked file and developed a hedonic price model for predicting premiums for specific plan designs.
Possible Health Plan Choices
Offered workers HMO Low-option PPO (PPO_Low $$) Medium-option PPO (PPO_Medium $$) High option PPO (PPO_High $$) Turn down offered insurance and do not
purchase individual HSA Individual HSA (HSA-Full Price) Employer-sponsored Health Reimbursement
Account (HRA) Employer-sponsored HSA (HSA-Shared
Premium)
Assigning health plans to MEPS individuals with no offer of employer coverage
1. Assumed all of these adults have a choice of 3 PPOs, a ‘stingy’ HSA plan, and uninsured.
2. Used MEPS linked insurance data to develop a hedonic price model to predict premiums for individual PPO plans, given specific plan attributes.
3. Used 2002 eHealthinsurance.com/HIAA survey statistics to scale individual PPO premiums by age.
4. Used eHealthinsurance.com for current premiums and cost-sharing attributes for HSA plans.
Possible Health Plan Choices
No Offer: Individual pays entire premium and the five plan choices are the same for everyone PPO_Low $$ PPO_Medium $$ PPO_High $$ HSA-Full Price UninsuredNote : All HSA plan offers include a $1,000 ‘starting account’ in the
premium amount. For example an HSA premium in the simulation of $4,000 includes $3,000 for the high deductible insurance and $1,000 for the health account.
Simulation Calibration Applied CDHP employer plan choice model
estimates to predict probabilities of plan choices for MEPS sample respondents
Model Calibrations % of adults who turn down employer offers
by income quartile % of adults in the individual market who are
uninsured by income quartile Applied national weights to the calibrated
model to represent the population 19-64, excluding full-time students, those enrolled in public insurance, and non-offered dependents with employer coverage through their spouse
Approximately 121.5 million adults
Policy Simulations Baseline take-up of HSAs from the Medicare Modernization Act
of 2003 Simulation (1): Bush Administration’s proposal
Refundable tax credit up to 90% of premium; maximum of $1000/adult, $500/child (up to two) Subsidy for singles with no dependents phased out at $30,000
adjusted gross income and $60,000 for families Simulation (2): Low income buy-in subsidy
Set HSA premium at $0 for < $15,000 annual income; 50% of premium for incomes between $15,000-$40,000; 75% of premium for incomes between $40,000 and $60,000.
Simulation (3): Full subsidy of HSA premium Set HSA premium to $0 for all, regardless of income.
Simulation (3a): Full subsidy of HSA premium for “generous” HSA policy
Simulation (4): Full subsidy of HSA premium for the non-working, non-public insurance population
-.3 to -.4 in the non-offered population (Marquis et al, 2004) – Ours are similar.
Take-up elasticity increases as income increases. This contrasts with the literature. Why? Probability of HSA take-up is positively correlated
to income (as opposed to an HMO, which is usually negatively correlated).
Implication is that lower income population need more inducement to take-up an HSA
Plan design matters. Greater take-up from a reduction in the donut
hole than an increase in the account size.
Policy Implications - 2 Other influential factors
Targeting the right population Need more targeting of subsidy – Want to avoid providing
too much subsidy to the group market. Implementation complexity
Timing of tax credit distribution – needs to be available at time of premium payment. Without this, the delay could be 12 months or later.
Want to avoid unnecessary administrative eligibility barriers like those experienced in the Trade Adjustment Assistance Reform Act of 2002
Supply side embrace Compared to Medicare Medical Savings Account in HIPAA
and other past policies, insurers are willing to build a new product.
Research Implications Donut can and should be modeled in plan choice
to get price response information. Health savings/reimbursement accounts can and
should be modeled in plan choice to see what is an appropriate amount – This will depend on objectives (current spending versus medical savings for later years).
Welfare implications for CDHP plans need to be considered in terms of short term account use and long term medical/retirement use – Need retiree information from employers to take next preliminary steps. Work could be used to contribute the health/retirement decision literature summarized by Gruber & Madrian, 1995.
Next Steps Refine the model to account for the following:
Individual-level vs. Household-level Income measurement issues for determining eligibility Non-offered adults who have access to employer
coverage through their spouse (~26 million adults) Offered but turned down coverage Premium inflation assumptions
Examine impact of other policy proposals and/or HSA plan designs on take-up.
Look at HSA take-up versus retirement saving choice.