intensity of health services and costs of care for previously uninsured medicare beneficiaries j....
TRANSCRIPT
INTENSITY OF HEALTH SERVICES AND COSTS OF CARE FOR PREVIOUSLY
UNINSURED MEDICARE BENEFICIARIES
J. Michael McWilliams, M.D.
Division of General MedicineBrigham and Women’s Hospital
Department of Health Care PolicyHarvard Medical School
June 26, 2006
Uninsured Near-Elderly Adults
• ~4 million and growing (Baby Boom)
• Difficulties acquiring coverage
• Greater risk of declines than younger adultsConsequences of uninsurance more severe
• Mortality: HR 1.43 uninsured vs. insured* HR 1.56 among adults with chronic conditions
• Proposals to expand coverage: Medicare buy-in
*McWilliams, Zaslavsky, Meara, & Ayanian. Health Aff 2004
Study Rationale
“The Committee hypothesizes that… uninsured persons gaining Medicare coverage at age 65 use health services more intensively and incur program costs higher than they would have had they been continuously insured prior to age 65… This question is one that merits further investigation.”
-- Committee on the Consequences of Uninsurance, Institute of Medicine 2004
Health and Retirement Study (HRS)
• 9,760 Adults ages 51-61 in 1992
• Biennial surveys through 2004
• Broad range of questionsHealth insurance coveragePhysician visits in prior 2 yearsHospital admissions in prior 2 yearsMedical expenditures in prior 2 years
Study Cohort:Near-Elderly with Chronic Conditions
• Hypertension (HTN), Diabetes (DM), Heart Disease (CHD), or Stroke (CVA)
• Excluded deaths and publicly insured
• At least age 66 by 2004
• Insurance status prior to age 65:Continuously Insured (all) n = 2090Ever Uninsured (1+y) n = 800Continuously Uninsured (3+y) n = 401
Statistical Analysis
• Multiple waves before/after 65 (age 59 to 69)
• Comparisons of utilization & expenditures:Before and after Medicare coverage at age 65Differential increases
• GLMs with log link function
• GEEs to account for correlated data
• Multiple Imputation for missing data
• Propensity score weighting to adjust for observed characteristics at age 59-60
Descriptive Comparisons: Unadjusted
Cont. Insured
Ever Uninsured
P-Value
Female (%) 53.8 57.1 <0.001
Race & Ethnicity
Black (%)
Hispanic (%)
5.7
3.2
12.2
11.8
<0.001
<0.001
Household Income
Lowest Decile (%) 4.9 19.2 <0.001
Education
<High School (%) 14.4 32.9 <0.001
Self-reported Health (1-5) 2.31 2.63 <0.001
Descriptive Comparisons: Adjusted
Cont. Insured
Ever Uninsured
P-Value
Female (%) 55.6 55.6 1.00
Race & Ethnicity
Black (%)
Hispanic (%)
9.3
6.6
9.3
6.6
1.00
1.00
Household Income
Lowest Decile (%) 11.1 11.1 1.00
Education
<High School (%) 23.8 23.8 1.00
Self-reported Health (1-5) 2.48 2.48 1.00
Hospital Stays: HTN, DM, CHD, CVAUnadjusted
0.2
0.3
0.4
0.5
0.6
59 61 63 65 67 69
Age
Hos
pita
l Sta
ys/2
y
Ever Uninsured Continuously Insured
Δ <65 -0.01 (P=0.66)
Δ >65 +0.14 (P<0.001)
Differential Increase +0.15 (P<0.001)
Hospital Stays: HTN, DM, CHD, CVAAdjusted
0.2
0.3
0.4
0.5
0.6
59 61 63 65 67 69
Age
Hos
pita
l Sta
ys/2
y
Ever Uninsured Continuously Insured
Δ <65 -0.04 (P=0.22)
Δ >65 +0.11 (P=0.01)
Differential Increase +0.15 (P=0.003)
Doctor Visits: HTN, DM, CHD, CVAAdjusted
7
9
11
13
15
59 61 63 65 67 69
Age
Vis
its/
2y
Continuously Uninsured Continuously Insured
Δ <65 -0.8
(P=0.04)
Δ >65 +1.8 (P=0.04)
Differential Increase +2.6 (P<0.001)
Total Medical Costs: HTN, DM, CHD, CVAAdjusted
17,000
24,300
15,800
30,800
10,200
37,000
5,000
10,000
15,000
20,000
25,000
30,000
35,000
40,000
<65 >65
$ (2
004)
/2y
Cont. Insured Ever Uninsured Cont. Uninsured
Δ <65 -6,800
(P=0.008)
Δ >65 +12,700 (P=0.10)
Differential Increase +19,500 (P=0.005)
Conclusions
• Significantly greater increases in health services and expenditures for previously uninsured adults gaining Medicare coverage
• Uninsured near-elderly adults w/ chronic diseases required more intensive and costly care as Medicare beneficiaries than similar adults who were previously insured
• Higher utilization after 65 exceeded lower utilization before 65
Policy Implications
• Substantial downstream reductions in services and costs of care
• Potential offsets to costs of coverage expansionTotal net costs of reformCosts to Medicare
• Importance of effective targetingSubsidies for Medicare buy-inExpand qualifying medical conditions