Cost effective choices
within different plan
designs
Judith H. Hibbard
Jessica Greene
Anna Dixon
Martin Tusler
University of Oregon
Funding provided by The Changes in Health Care Financing and Organization
(HCFO) initiative, a program of The Robert Wood Johnson Foundation
Will CDHPs increase cost effective
choices?

Use of health information tools will help
consumers make more cost effective
choices. They can help them determine:
 When
care is actually needed
 What kind of care is needed
 When self-management strategies are
appropriate, and which ones to use
 When no action is necessary
Operationalizing “Cost-Effective
Choices”: Measuring “unproductive
utilization”



Using a schema based on the Oregon Health
Plan priority list, ICD-9 codes are used to
classify office visits into productive (evidence for
efficacy) vs unproductive care (little or no
evidence of efficacy)
Acute and chronic visits are examined
Ratios of acute “unproductive” visits are
calculated by dividing the number of
“unproductive” acute visits by the number of all
acute visits (for those with at least 3 visits in
2003) - the ratios indicate a “pattern of use”
Examples of the most common
“unproductive” acute and chronic visits

Acute
 Acute
non-specific upper respiratory infection
 Cough
 Acute pharyngitis
 Impacted cerumen (earwax)

Chronic
 Allergic
Rhinitis
 Rotator cuff syndrome non-specific
Claims data for employees was
included in the analysis



Examined whether there was a change in
“unproductive care” chronic or acute care from
2003 to 2004 (year employees entered CDHP)
Examined differences across the three plans
looking at differences in ratios of “unproductive
acute and chronic care
Analysis controls for 2003 ratio, employee level,
gender, age and a co-morbidity index
Ratios of “unproductive” acute visits
For those with a minimum 3 visits in 2003
High
Deductible
CDHP
Lower
Deductible
CDHP
PPO
2003
.50
.46
.48
2004***
.50
.42
.47
Key: *p<.05, *** p<.001
Ratios of “unproductive” chronic care
visits
For those with a minimum 3 visits in 2003
High
Deductible
CDHP
Lower
Deductible
CDHP
PPO
2003
.41
.41
.42
2004
.39
.35
.39
None of the results were significant
Use of “unproductive” acute care visits
in 2004
Beta weights from regression
PPO compared to
lower deductible
CDHP
PPO compared to
high deductible
CDHP
High deductible
CDHP compared to
lower deductible
CDHP
.108***
N.S.
.044*
Controlling for unproductive acute care in 2003, age,
gender, employment type, and Charlson Co-morbidity Index
Key: *p<.05, *** p<.001 NS not significant
Use of “unproductive” chronic care
visits in 2004
Beta weights from regression
PPO compared to
lower deductible
CDHP
PPO compared to
high deductible
CDHP
High deductible
CDHP compared to
lower deductible
CDHP
.034***
N.S.
.044*
Controlling for unproductive acute care in 2003, age,
gender, employment type, and Charlson Co-morbidity Index
Key: *p<.05, *** p<.001 NS not significant
Evidence for cost-effective choices
in the claims data



This is only true for the lower deductible CDHP
enrollees
Lower deductible enrollees changed patterns of
use - to use fewer unproductive visits
High deductible enrollees did not differ from the
PPO in their unproductive utilization patterns.
Their patterns were largely unchanged from preenrollment period.
Conclusions


Evidence supports both the hopes and the fears
about CDHPs
Study design allows us to:
 hold
constant cost exposure and look at the impact of
the CDHP plan elements (comparing PPO enrollees
to lower deductible CDHP enrollees)
 Hold constant CDHP plan elements and examine the
impact of greater cost exposure (comparing high and
lower deductible CDHP enrollees)

The basic theory underlying the CDHP approach
seems to be correct, but greater cost sharing
appears to be counter productive
Conclusions continued

Limitations and future studies
 Cannot
rule out outcomes a result of selection
 Need to replicate findings beyond one employer



Future studies should examine different levels of
cost exposure and calibrate effects
Who are the people that do well in a CDHP?
What are their characteristics?
How well do people with less education and
income manage in a CDHP?