Insurance Product Design and Its Effects: Tradeoffs along the Managed Care Continuum

1
Insurance Product Design and Its
EffectsTrade-offs along the Managed Care
Continuum
Presentation at the University of
Minnesota February 25, 2002

• Peter Kemper
• Ha T. Tu
• James D. Reschovsky
• Elizabeth Schaefer

2
Motivation

• Public concern about managed care
• Policy interest in regulating managed care
• Purchaser interest in providing choice
• Insurance product designs have proliferated
• Must look beyond the managed care label
• Question How does insurance product design
  affect service use, access, and consumer
  assessments of care?

3
Previous Literature

• Many studies on effect of HMOs, reviewed by
• Miller and Luft (1994, 1997)
• Glied (2000)
• Vary greatly in representativeness, controls
• Much of it from earlier era of managed care
• Very little beyond HMO/non-HMO comparison
• This study builds on authors previous work

4
Framework Managed Care Continuum

• Insurance products are combinations of features
  to encourage efficient resource use
• Patient cost-sharing
• Provider care management tools
  (e.g., provider networks, gatekeeping,
  capitation)
• Continuum unmanaged to heavily-managed as more
  care management tools are added

5
Framework Analyze Product Types

• Analytic challenge high correlation because
  features are bundled together systematically
• Some features work together
  (e.g., capitation virtually implies
  gatekeeper)
• Trade-off between patient cost-sharing and
  provider care management tools
• Constructed types based on 4 important features
• Important factors
• Correlated with other features

6
Stylized Features of Product Types
7
Other Product Features(percent)
8
Data

• Community Tracking Study Household Survey
• Telephone survey
• Family respondent plus self-response module
• Conducted in 1996-7
• CTS Insurance Followback Survey
• Contacted health plans identified by family
• Information on insurance features
• Linked to individuals based on product

9
Sample

• 60 sites selected randomly
• Metropolitan statistical areas
• Stratified geographically and by size
• Sampling proportional to population
• All adults and randomly selected child
• Analysis sample individuals with
  employer-sponsored insurance (n25,560)

10
Model
Y ? I ? X ? M ? Y Service use, access,
assessment I Insurance product X
Individual characteristics M Dummy variables
for 60 markets
11
Individual Characteristics

• Health status (SF-12)
• Income
• Demographics
• Family size
• Willingness to trade cost for provider choice
• Whether offered choice of insurance

12
Analytic Methods

• Logit models for dichotomous variables
• Two-part model for continuous variables
• Predictions reported for each product type
• Standardize for differences in enrollees health
  status, income, preferences, etc.
• F-test for equality across all types
• Accounted for complex sample design

13
Overview of Findings

• Results tend to follow continuum
• Service use Differences in use of primary and
  specialty care but not costly services
• Access No differences overall but trade-off
  between financial and administrative barriers
• Consumer assessments Become less favorable
  under heavily-managed products

14
Total Ambulatory Visits (per 100 persons)
407
407
396
378
369
37
35
63
28
29
285
253
246
264
239
101
104
97
85
91
FFS
15
Other Service Use Differences Are Not Significant

• Hospitals
• Surgery
• Emergency Departments

16
Unmet Need or Delayed Care (not significant)

FFS
17
Reason for Unmet Need, Delayed Care
Financial barriers
Administrative barriers

18
Family Out-of-Pocket Cost
FFS
19
Less than Very Satisfied with Health Care
42.8

39.8
36.3
36.4
35.7
10.0
9.1
8.4
7.9
6.7
FFS
20
Less than Very Satisfied with Choice of Physicians
Specialists
Primary Care Physicians
37.5


34.4
30.8
30.2
29.6
28.8
27.3
26.3
25.3
23.7
9.7
9.3
9.0
8.9
8.6
7.8
7.2
6.6
6.0
5.4
FFS
FFS
21
Thoroughness of Examination Less than
Excellent(primary care physician visits)

69.0
67.0
65.3
63.7
62.6
12.2
9.9
8.7
7.6
7.7
FFS
22
Less than Strong Trust in Physician
Refers to Specialist When Needed
Puts Medical Needs First
41.7


40.7
36.5
34.9
34.2
34.0
30.9
30.3
29.5
29.3
20.2
16.5
16.1
13.7
11.7
11.0
9.2
7.1
7.1
5.9
FFS
FFS
23
Implications for Research

• Need to look beyond the labels
• Difficult to separate effects of features
• Way products are characterized matters
• More research on characterization is needed

24
Implications for Offering Choice

• Choice may not eliminate low ratings
• Trade-off cost vs. administrative barriers
• Choose low cost but still assess unfavorably

25
Implications for Regulatory Policy

• Beware of risk of regulating specific care tools
• Difficulty of separating effects of tools
• Regulate options that some prefer out of
  existence
• Move to managed care lite mitigating backlash?
• Address service delivery in entire system
• Low ratings for all product types
• Challenges throughout health system