Impact of a Value-based

Formulary on Medication
Utilization, Health Services
Utilization, and Expenditures
Kai Yeung, PharmD, PhD, Anirban Basu, PhD, Ryan N. Hansen, PharmD, PhD, John B. Watkins,
PharmD, MPH, BCPS, and Sean D. Sullivan, PhD
Medical Care. (2017). 55(2): 191-198

1
Outline
Introduction
Methods
Results
Discussions
Critiques

2
INTRODUCTION
n 1.49
copayment
n cost sharing

n Value-based insurance (VBID) plans

treatment

1. evidence-based data plans
2. Health benefit planspreventive care, wellness visits,

3. unnecessaryhigh cost-sharing
copayment

3
INTRODUCTION
n
()

1. (1.5%-9.4%)
2.
1) Lower member (e.g. out of pocket) medical expenditures
nonmedical expenditurestotal expenditures
2) Increases health plan medication expenditures
3) No change
limitation:
1.

4
INTRODUCTION
n In 2010, Premera Blue Cross implement
Value-based formulary(VBF) benefit among its employees and
dependents
cost-effectiveness analysis (CEA) copayment
clinical, economic, bioethical experts CEA
estimatesICERcopayment tier
1. Higher ICER: higher copayment tiers
2. Lower ICER: lower copayment tiers (to incentivize use)

5
INTRODUCTION
n Value-based formulary(VBF)
1. Higher ICER: higher copayment tiers
2. Lower ICER: lower copayment tiers (to incentivize use)

Special case: drugs that had additional value not

reflected by their ICER. These values include ethical
issues, disease rarity, unmet clinical needs, regulatory
requirements, and other societal considerations. 6
INTRODUCTION
n However,
Unclear whether the available CEA evidence was sufficient to appropriately
estimate the value of drugs.
Copayment tier assignment is based on population average cost-
effectiveness estimation, but cost-effectiveness is patient specific.
Better health may come at a higher cost (total health expenditure).

Purpose
Investigate the impact of a VBF on medication and health services utilization and
on medication and nonmedication expenditures from member, health plan, and
member plus health plan (overall) perspectives.
7
METHODS

Sample, data
Study design
source

Measurement, Statistical
outcomes Analyses

8
 METHODS

Sample, data source

nIntervention group nControl group

The population of employees and
dependents.
Aged 0-64
Covered by PPO employer-sponsored
plans administrated by Premera Blue
Cross (the largest private health plan
in Washington State).
Enrolled at least 1 year before VBF
implemented.
The population of employees and
dependents.
Aged 0-64
Covered by 5 Premera employer-
sponsored plans without any changes.
Similarity: industry classification,
geography of residence, medication
copayment tiers.

9
METHODS

Measurement, outcomes
n Demographics:
(age, sex, ZIP code of residence,
relationship to employee)
n Expenditures:
(amount paid by member and
amount paid by health plan)
n Plan characteristics:
(benefit renewal month and
medical benefit relativity value)
n Prescriptions fills:
(code number, number of days
supply, date dispensed, place of
purchase)
n Nonmedication services:
(date of service, place of service,
length of hospitalization, procedure,
diagnosis, and revenue codes)

Medical benefit relativity value:

An index of medical benefit from insurance plan.
The values range between 0 and 1.
A value of 0.75 means that a health plan pays 75% of medical expenses. 10
METHODS

Measurement, outcomes
n Medication: The study measured the per member per month (PMPM)
1. Probability of filling.
2. The days supply of the medication.
n Health services utilization: (PMPM)
1. Probability of ED visits, hospitalization, office visits.
2. The number of ED visits, hospitalization, office visits.
n Three perspectives of expenditures:
member, health plan, and overall medication and nonmedication expenditures
PMPM.

11
METHODS

Measurement, outcomes
n Secondary analyses- assessed medication utilization based on 2
categorization:
1. Medications moved into lower copayment tiers, higher copayment tiers, or
no change in tier in the VBF.
2. Medications moved into the preventive tier or into tiers 14 in the VBF.

n Falsification test
Assessed the expenditures for vision services

12
METHODS

Study design
July 2006 - June 2013

VBF
Late
Early pre-VBF pre-VBF post-VBF

2006 2009 2013

2010

Control for all benefit changes other than the implementation of the
VBF policy in the statistical analyses. 13
 METHODS

Statistical Analyses
n Confirmed the similarity of the control group to the intervention group in pre-VBF
outcomes trends.
n Generalized estimating equations with 2-part models.
n Medication & utilization Binomial distribution with logit link to model the
probabilities. Poisson distribution with log link to model the counts.
n Medication & non-medication expenditures Binomial distribution with logit link
to model the probabilities. Gamma distribution with logit link to model
expenditures.
n Adjusted for
n Individual-level characteristics (sex, age, total health care expenditure>$100,000)
n ZIP code level characteristics (household income, urban residence,...)
n Plan-level characteristics (medical benefits relativity value)
n Fixed effects for calendar (January December)
n Study period (early pre-VBF period, late pre- VBF period, and post-VBF periods)

n STATA version13.1 14
 RESULTS
1. Population characteristics

n Two groups were similar in demographic and socioeconomic characteristics in the

pre-policy period
RESULTS
2. Changes in Mean Copayments Due to the VBF

68%
28%
4%

16
RESULTS
3. Changes in Utilization and Expenditures after Adjusting for Secular Trends

n The VBF policy had no statistically significant overall impact on

medication utilization.

n (copayment tier)tier
utilization(0.02 PMPM probability of
fill, 11%; P<0.001)

n preventive tierutilization
(0.02 PMPM probability of fill , 13%; P<0.001)
n no statistically significant changes in
probability or quantity of use for ED visits, hospitalization, or
office visits
17
RESULTS
3. Changes in Utilization and Expenditures after Adjusting for Secular Trends

n Member medication expenditures increased significantly by $2 PMPM

(9%; P = 0.004).
n Health plan and overall medication expenditures decreased significantly
by $10 PMPM (16%; P=0.02) and $8 PMPM (10%; P = 0.01).
n No statistically significant impact on member/health plan and overall
nonmedication expenditures.
18
 RESULTS
3. Changes in Utilization and Expenditures after Adjusting for Secular Trends

Observed and expected medication expenditures PMPM in intervention (VBF) and

control groups combining expenditures from member and health plan perspectives. 19
DISSCUSSIONS
n (medication expenditure)10%
($8 PMPM)
n overall medication utilization or health services
utilization or nonmedication expenditures.
n health plan member.
n 3post-policynet savings overall
medication savings $1.1 million.
n (medication utilization analysis) VBF
higher value drugs
n employer-sponsored health plan
net savingslower insurance
premiumsmember
20
DISSCUSSIONS
n Policy Implications
VBF may be a nuanced way to cost-share such that patients are
shifted toward higher value drugs and other health services
utilization are minimized.
Future iterations of VBID should consider an explicit estimate of
value in Policy Implications. (CMS)
n Limitations
working-age population and their dependents
poor, elderly, or chronically ill
Unobserved confounders: affected these trends differentially
during the post-policy period.
Did not know the impact of the VBF on actual health outcomes.

21
CRITIQUES
n
1. copayment,
tierstier
2. (member, health plan, and overall)
2015Premera Blue Cross
VBFDesign, implementation, and first-year outcomes
3. follow-up
4.

22
CRITIQUES
n
1.



2. tiers

3. member, health plan

23
CRITIQUES
n
4. VBFPMPM
matching

24
THANK YOU.

25