The Congressional Budget Office analyzed how changes in prescription drug use affect Medicare spending on medical services. It found that higher drug use leads to lower medical costs, as drugs help control diseases and reduce complications. The CBO reviewed studies on this "offset effect" and found that a 1% increase in drug use leads to a 0.2% decrease in other medical spending. As an example, the CBO estimated that closing the Medicare Part D coverage gap would increase drug use by 5% and lower medical costs by 1%, reducing the total cost of closing the gap.

1. Congressional Budget Office
Presentation to the AcademyHealth Annual Research Meeting
Melinda B. Buntin, Ph.D.
Deputy Assistant Director for Health
Tamara B. Hayford, Ph.D.
Principal Analyst
This presentation provides information published in Offsetting Effects of Prescription Drug Use on Medicare’s Spending
for Medical Services (November 2012). See http://www.cbo.gov/publication/43741
June 23, 2013
Offsetting Effects of Prescription Drug Use
on Medicare’s Spending for Medical Services

2. C O N G R E S S I O N A L B U D G E T O F F I C E
Overview
■ Why did CBO revisit the effect of changes in drug use on
medical service use?
■ How does the use of prescription drugs affect medical
spending?
■ Methodology and methodological issues to consider
■ Literature and results
■ Example of applying the medical-drug “offset”

3. C O N G R E S S I O N A L B U D G E T O F F I C E
Why Revisit the Relationship Between Drug Use
and Medical Spending?
• New evidence, advances in the literature
• Congressional interest in a range of drug cost-
sharing proposals, e.g.:
• Lower LIS generic copay and increase LIS brand copay
• Increase manufacturer’s discount in the coverage gap
• Fill the donut hole

4. C O N G R E S S I O N A L B U D G E T O F F I C E
Higher
cost to
enrollee
Less controlled
disease
• More complications
• More ER visits
• More hospitalizations
• Etc.
How a Drug Price Affects Medical Spending
Reduced Rx
Use
Reduced Rx
adherence
Reduced Rx
continuation
Might vary by…
Health of affected group
Size of change in Rx price
Direction of change in price
Lower Rx
Spending

5. C O N G R E S S I O N A L B U D G E T O F F I C E
Methodology:
Study Selection and Results Calibration
• Select studies:
• Analysis of changes in overall drug use
• Populations “applicable to” Medicare population
• Calibrate study results:
• Adjust to be consistent with overall Medicare population
• Adjust to be consistent with overall medical spending
• Scale to be consistent with a 1 percent change in
prescription drug use
• Calculate average of calibrated, scaled results weighted by
study quality

6. C O N G R E S S I O N A L B U D G E T O F F I C E
Issues to Consider
• Are effects symmetric to increases and decreases?
Linear?
• Can the factor be applied to subgroups of the Part D
population?
• Can we apply the logic to specific drug classes (e.g.
biologics)?

7. C O N G R E S S I O N A L B U D G E T O F F I C E
• One study estimated the impact of pharmaceutical
policies on a broad population outside of Medicare
(Gaynor, Li, & Vogt, 2007)
• Four studies estimated the impact of pharmaceutical
policies on Medicare beneficiaries prior to Part D
implementation (Chandra et al., 2010; Stuart et al.,
2009; Shang and Goldman, 2007; Hsu et al. 2006)
• Three studies compared medical expenditures before
and after Part D implementation (McWilliams et al.
2011; Afendulis et al. 2001; Zhang et al. 2009)
Estimates of the Size of the Offset

8. C O N G R E S S I O N A L B U D G E T O F F I C E
Effect of a 1 Percent Increase in Drug Use on Other
Medical Spending
Scaled Results Range -2/3 % to +1/3%
Scaled Range Excluding Max and Min -4/10% to -1/10%
Weighted Average -1/5 %

9. C O N G R E S S I O N A L B U D G E T O F F I C E
The relationship between changes in drug utilization and non-drug
medical spending appears to be symmetric and linear
-12
-10
-8
-6
-4
-2
0
2
4
6
8
-30 -20 -10 0 10 20 30
Calibrated%changeinnon-drugmedicalspending
% change in Rx utilization
Zhang et al (2009)
outlier

10. C O N G R E S S I O N A L B U D G E T O F F I C E
Applying the Offset
■ When?
– For policy changes that are estimated to change the quantity of drugs
consumed in the Medicare program
– Not for policy changes that affect other programs
– Not for policy changes that would not directly induce a change in the
quantity of drugs consumed
■ How?
– First: estimate a proposal’s direct effect on prescription drug costs
– Next: estimate the effect on the number of prescriptions filled
– Last: calculate any resulting offsetting effect on spending for medical
services.

11. C O N G R E S S I O N A L B U D G E T O F F I C E
Applicability to Population and Drug Subgroups
■ Population subgroups:
– CBO will apply the factor to policies affecting broad
subgroups (e.g., LIS or non-LIS population)
– Narrower populations decided on a case-by-case basis
(Study results may or may not be applicable)
■ Drug subgroups:
– Studies addressed changes in overall drug use
– Effects within therapeutic classes will require further
literature review for those classes

12. C O N G R E S S I O N A L B U D G E T O F F I C E
Example: Closing the Part D Coverage Gap
■ Components of closing the gap:
– 50% manufacturer’s discount on brand drugs in the coverage gap (for
non-LIS beneficiaries only) began in 2011.
– Part D Plans gradually responsible for more coverage until, in 2020,
they pay for 25% of brand drugs & 75% of generics.
■ Estimated effect:
– Total consumption of drugs by non-LIS will increase by 5% by 2018.
By 2018, spending for medical services expected to fall by 1% for this
population.
– For 2013-2022 period, Medicare medical spending estimated to fall by
$35 billion (out of $5.6 trillion).
Estimated cost of closing the gap for 2013-2022 period drops from $86
billion to $51 billion.