University of Southern California Dissertations and Theses

(USC Thesis Other)

US biosimilar market entry and uptake: US commercial payer preferences in access decision making

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

ELIZABETH LYDIA SCHWARTZ
DECEMBER 2015
UNIVERSITY OF SOUTHERN CALIFORNIA
FACULTY OF THE USC GRADUATE SCHOOL
PhD
(PHARMACUETICAL ECONOMICS AND POLICY)

US Biosimilar Market
Entry and Uptake
US Commercial Payer Preferences in Access Decision Making
1

TABLE OF CONTENTS
Page
Chapter 1: Introduction and Background 3
I. US Health Care Costs and Contribution of Biologic Spend 3
II. Access Concerns and Opportunities 4
III. Biosimilars: Explanation of Barriers and Associated Challenges 5
IV. Perspectives and Novel Policy Considerations 9

a. Innovator and Biosimilar Manufacturer Interests 9

b. Payers: Private and Government 10
V. Research Objective and Explorative Insights 13
VI. Refined Research Question and Project Objective 16
Chapter 2: Literature Review 18
I. Background on Choice Behavior 18
II. Application of Discrete Choice Methods in Healthcare Valuation 19
III. Discrete Choice Methods in Healthcare Decision-Making 20
a. Opportunities for Research Expansion 22
IV. Use of Organizational Decision Making Methods in other Economic Sectors 23
V. Organizational Decision Making Perspectives in Policy Adoption 24
Chapter 3: Study Design and Methods 27
I. Survey Development 27
a. Product Attributes and Proposed Levels 27
b. Sequence of Experiments 31
c. Profile Generation 32
d. Attribute Ranking Exercise 32
II. Hypotheses 33
III. Sample 34
IV. Model Specification and Estimation 35
2

Chapter 4: Results 39
I. Survey Sample 39
II. Qualitative Results 40
a. Biosimilars compared to small molecules 40
b. Awareness and effect of FDA regulation 41
c. Formulary review and decision-making process for biosimilars 43
d. Future benefit design 45
III. Quantitative Results 46
a. Product Choice Selection 46
b. Formulary Decision Making 48
i. Biosimilar at Parity vs. Biosimilar Non-formulary 49
ii. Biosimilar Preferred for New Patients vs. Biosimilar Non-formulary 51
iii. Biosimilar Preferred for All Patients vs. Biosimilar Non-formulary 53
IV. Attribute Ranking Exercise 54
Chapter 5: Discussion 57
I. Primary Hypotheses 57
II. Exploratory Hypothesis 60
III. Additional Payer Policy Considerations 62
IV. Research Limitations 63
a. Methodological Limitations 63
b. Applicability Limitations 64
V. Future Research 66
VI. Conclusion 68
References 70
Appendix A: Instrument Key 79
Appendix B: Example of Two-step Discrete Choice Example 80

3

Chapter 1: Introduction and Background

This chapter will provide a background on healthcare spending in the
United States (US) focused specifically on the increasing drug expenditure for
targeted, biologic therapies. Given the mounting budgetary pressure, this section will
explore drug access concerns and discuss regulatory pathways for biosimilar versions of
these drugs. In addition to describing the market dynamics surrounding both the supply
and the demand for biosimilars drug utilization and management policy perspectives will
be presented and managed care decision-making trade-offs considered. The chapter will
conclude by outlining the primary objective of the research and exploring pilot study
insights that serve as the basis of the survey instrument used to elicit managed care
preferences.
I. US Health Care Costs and Contribution of Biologic Spend
Well-documented and characterized, the rate of growth in total healthcare
expenditures in the US continues to rise over time. In 2013, healthcare expenditures
accounted for $8700 per capita, totaling 16.4% of the gross domestic product (GDP) (1).
Surpassing all countries globally and with upward projections in the future, dominating
healthcare costs in the US in recent years comprise a target sector of the economy
where concern and interests span private industries, government, public policy agencies,
and healthcare delivery systems. Dissecting the drivers accounting for the escalation in
total expenditures requires a focused analysis of individual spending by healthcare
segments as well as an understanding of the associated dynamic trends. Considered a
market leader in prescription drug innovation in the absence of regulated drug price
controls, prescription drug spending in the US represents a convoluted market of interest
with multiple moving parts driving growth.
4

Since 2007, healthcare cost growth has slowed down due to the low growth rates
of prescription drug spending. In 2012, there was a decrease of drug spending by 1%
followed by an increase to 3.2% in 2013, which is still quite modest compared to growth
rates in previous years. The largest contributing factor of these decreasing growth rates
is the expiry of patents for small molecule generics (2). Although prescription drug
expenditures and utilization for small molecule drugs have been stabilizing, the
therapeutic advancement and continued development of targeted biologic therapies has
led to these drugs accounting for 28% of all prescription medications with increased
growth of 9.6% from 2012 to 2013 (compared to 0.1% for small molecules). (2).
Biologic drugs do offer significant clinical benefits and improvements in disease
management in areas such as rheumatoid arthritis and oncology; however, these agents
have higher costs compared to small molecule treatments partially due to the complexity
of development and manufacturing (3). With an increase in the development of biologic
drugs delivering superior clinical outcomes, clinicians, payers, and patients alike are
balancing the clinical benefit with the additional costs of these more advanced, targeted
treatment options.

II. Access concerns and opportunities
Implications closely correlated with the high cost of biologic care lead to
questions about patient access to these novel therapies. Both private and government
payers bear the burden of reimbursement for biologic treatments. In managing the
benefit budget, payers must balance adequate coverage that optimizes health benefits
with the allocation of limited financial resources among all beneficiaries. However, in the
case of biologic drugs which are significantly more expensive, there is greater impetus to
manage the budget for these agents more closely compared to that for branded small
molecules. To ensure appropriate usage of biologics and alleviate budgetary
5

constraints, managed care payers have instituted utilization mechanisms, such as high
cost-sharing and prior authorization requirements, to manage access to biologic drugs
(4-6). Evidence suggests that level of benefit generosity and variability in access to
biologic categories among managed care organizations (MCOs) largely influence patient
use and adherence to therapy (7). For example, benefit policies limiting access by
instituting high co-payments affect cost-sensitive patients by increasing their likelihood to
exhibit non-compliant behavior. Thus, these policies and outcomes may have
unintended consequences, leading to either unrecognized therapeutic benefit or
suboptimal disease management and additional downstream medical costs to payers
and society (8).
In addressing appropriate and equitable access to high-cost biologic care,
management policies for existing branded drugs are not the only mechanisms that can
leverage access. With the approaching expiration of patents for many top-spend and
highly utilized biologic products, health care policy experts and FDA regulators have
deliberated over the feasibility and methods for treating biologics like small molecule
pharmaceuticals with regards to generic availability. Analogous to the introduction of
generics for branded, small molecule chemical entities, payers and patients are
interested in the prospective approval of biogenerics or more commonly referenced,
“biosimilars.” The anticipated cost-savings, which can mitigate treatment costs and
potentially expand access, serve as primary drivers spurring the interest of payers and
policymakers in the development of biosimilars. Generic availability for many small
molecule drugs has led to a relatively rapid conversion from brands in crowded drug
categories as well as associated reductions in drug costs and average patient co-
payments (9). Given the advantages afforded by the availability of generic small
molecules, there is mounting interest among policy makers as to whether both payers
6

and patients can reap comparable benefits by using similar but less expensive product
substitutes for biologics.

III. Biosimilars: Explanation of Barriers and Associated Challenges
Despite the opportunities for availability, there is much debate surrounding the
regulation of biosimilars which is rooted in both legal and public health concerns.
Contrary to chemical drugs which are synthesized in a fixed chemical process, these
biologic medicines consist of large, complex molecules produced by recombinant DNA
technology. Due to the more sensitive molecular structure of these proteins, the greater
index of variability between innovator products and biosimilars can lead to differences in
clinical efficacy and safety. The challenge in determining the significance of these
dissimilarities is that with subtle changes in molecular folding and subsequent biological
activity, current analytical methods may be unable to adequately characterize these
differences.
Compounding safety and efficacy concerns, the development and manufacturing
of biosimilars pose both regulatory barriers and important economic concerns. In the
US, guidance by the FDA up until recently was restricted to the assessment of copies of
biologic drugs. This was due to the lack of authority to establish a framework for
approval through an abbreviated pathway. For small molecule drugs, abbreviated filing
for generic approval is regulated under the 1984 Drug Price Competition and Patent
Term Restoration Act (Hatch-Waxman). Through this pathway, generic manufacturers
may file for FDA approval upon patent expiration of the brand while relying wholly or in
part upon the innovator’s safety and efficacy data. Through these mechanisms, the
abbreviated pathway can encourage market competition for patent-expired drugs by
lowering the cost of entry into the market.
7

In 2010, the Patient Protection and Affordable Care Act (PPACA) passed by
Congress authorized the FDA to provide guidance for developing an abbreviated
pathway for biologic drugs. The legislation distinguishes pathway designations for
biosimilar products and a more stringent designation for “interchangeability.” Biosimilar
products require data showing that their structure is “highly similar” to the innovator
product and exhibits no clinically meaningful differences relative to the innovator in
safety, purity, and potency. Unlike small molecules with therapeutic substitutability
ratings, biosimilars and their corresponding innovator products may exhibit small
structural differences. Requiring a higher standard of evidence than biosimilars,
interchangeable products necessitate additional data showing that the biologic copy can
be expected to produce the “same clinical result in any given patient” compared to the
innovator product. In this case, there must be no demonstrable risk in switching
between the innovator and interchangeable biosimilar, and the two are said to have the
same active ingredient (10).
Although the legislation communicates clear definitions, the FDA continues to
engage in ongoing discussions regarding details of the pathway mechanisms and
execution of the legislation. The implications of policies concerning how the FDA
regulates biosimilars and interchangeable products have fostered much debate and
hesitation in developing an abbreviated pathway. Questions about substitutable
interchangeability of therapy, variation in state authority in mandating generic
substitution, and patient safety are important regulatory considerations for
implementation.
In the US, the FDA has approved one biosimilar to date. Approved on March 6,
2015, Zarxio is a biosimilar to Amgen Inc.’s Neupogen (filgrastim) and is produced by
Sandoz, a generic division of Novartis. Zarxio is a granulocyte colony-stimulating factor
(G-CSF) used to treat neutropenia (characterized by an abnormally low white blood cell
8

count). Zarxio is licensed for the same indications as Neupogen and is classified without
interchangeable status. Despite regulatory approval, however, Zarxio is not
commercially available in the US due to litigation issued by Amgen regarding compliance
with the FDA pathway guidance for approval and patent infringement. Without promotion
and commercial availability, managed care payer decisions for Zarxio have not yet been
determined nor have any changes to coverage impacted current patient care.
Along with the approval of Zarxio, the FDA has published guidance on the
classification and listing (“purple book”) that will be used to identify approved biosimilars
and their corresponding reference products. The “purple book” includes and will include
a list of all biological products including biosimilar and interchangeable products that are
licensed by the FDA. However, the FDA has not finalized guidance on the naming
convention for biosimilars and interchangeable products as the regulatory authority is
still considering how or whether a proper unique identification for each biosimilar is
necessary and sufficient to ensure appropriate product safety monitoring.
Closely related to regulatory barriers in pathway development, barriers to market
entry and patient access also exist. Previous research in economic modeling of
biologics indicates that biosimilar list prices will not exhibit the same margin of difference
to innovator biologics as generic pricing for small molecule brands. The economic
literature suggests that difference in pricing between innovator biologics and biosimilars
(i.e., ratio of generic to branded price) will be significantly larger than that observed with
chemical branded entities compared to their generic counterparts (11). One driver of
this variation is the complex manufacturing processes required for biologic development
which entail higher fixed costs compared to small molecule drugs. With a higher cost for
market entry, fewer biosimilar manufacturers will seek approval unless a feasible
abbreviated pathway can assist in mitigating the costs for research and development.
Furthermore, estimates of the number of entrants and price of biosimilar products are
9

related to market size as a predictor of expected profitability. According to modeling
studies, the number of competitors is estimated to range on average from 2 to 5 while
the ratio of prices for biosimilars to innovators ranges from 65% to 90% depending on
the market size for a given therapeutic class. By comparison, economic models of small
molecule drugs over time have reported significantly smaller generic-to-brand price
ratios, with generic prices ranging on average from 53% of the branded price after the
first year to 30% of the branded price in the long term (12). With modest cost offsets
and biosimilar prices expected to approximate those for innovator products, access and
market conversion to these products may be more limited than in the case of small
molecule generics. With the differences in regulation and the perception of biosimilar
safety and efficacy compared to small molecule generics, cost structure differences
contribute to market environments where payers face novel policy decisions. Therefore,
pending the approval and commercial availability of biosimilars, it is imperative to
research, analyze, and understand the extent to which biosimilars can contribute
meaningful value with cost savings for society.

IV. Perspectives and Novel Policy Considerations
a. Innovator and Biosimilar Manufacturer Interests
As with the introduction of generic competition, innovator biotechnology firms are
concerned about the impact of biosimilars on market share erosion. The Biologics Price
Competition and Innovation Act (BPCIA) within the Patent Protection and Affordable
Care Act (PPACA) provides patent protection guarding market share and profits for a
specified period of time for both innovator small molecule and biologic therapies. The
BPCIA specifies a 12-year data exclusivity period for innovator biologic manufacturers.
In addition to patent protection, data exclusivity guards the product data and testing files
of the innovator manufacturer after FDA approval and prevents biosimilar manufacturers
10

from filing for an abbreviated pathway approval until the period has lapsed. As an
expression of trade secrets, data exclusivity provides additional protection to not just
product patent rights but the manufacturing processes and clinical trial data as well.
However, the length of this period is highly debatable as it is closely related to the trade-
off between innovation and competition. While innovator firms prefer longer periods of
data exclusivity, biosimilar manufacturers prefer shorter periods for expediting time to
market entry and expanding market share opportunities for growth. Unlike the small
molecule competitive landscape, biologic competition and abbreviated filing impose
additional hurdles to market entry. Policymakers continue to evaluate how specification
of the data exclusivity period can lead to downstream market shifts and consumer
access. The timing of market entry for biosimilars and the downstream effects of
delayed competition may foster payer preferences for managing biosimilars which differ
from those for managing familiar small molecule generics.
b. Payers: Private and Government
Currently, health care expenditure funding in the US is distributed almost equally
between private and government payers; estimates in 2013 showed the breakdown to
be approximately 48.2% public and 51.8% private (1). In addition to high-level policy
decisions concerning the regulation of biosimilars and economic dynamics, payers are
primarily responsible for determining how these products can be integrated into health
care delivery systems on a micro level. With safety and efficacy concerns leading to
differences in regulatory guidance and price referencing expectations, value perceptions
concerning biosimilars may differ significantly from those for small molecule generics.
Furthermore, different decision-making preferences may imply that the policies outlining
biosimilar coverage and reimbursement may also differ from those currently used for
managing generic small molecule drugs. Understanding the trade-offs influencing
decision-making for patient access is critical to projecting and forecasting the potential
11

budgetary relief offered by biosimilars. Despite the differences between managing small
molecule generics and biosimilars and the potential influence on uptake, cost-analyses
can provide payers with a baseline for budget expectations. In 2007, the Congressional
Budget Office (CBO) estimated the potential cost-offsets expected with implementing the
abbreviated pathway approval process for biosimilars. The results of the CBO model
showed that over the period from 2009 to 2018, biosimilars could reduce total healthcare
expenditures by $25 billion, which represented 0.5% of the national spending on
pharmaceutical drugs (13). Considering the modest projected cost-offset when
evaluated in aggregate, research is needed to understand how payers evaluate cost
differences and how their assessments affect decision-making concerning access to
therapy in their respective patient populations.
Anticipated cost-savings gained from the entry of biosimilars into the market is a
fundamental interest of both private and government payers; however, variations in the
mechanisms for management and the representative patient populations can influence
how decision makers perceive the value of biosimilars. For private commercial payers,
formulary placement and benefit coverage vary greatly depending on benefit coverage
design (ie, preferred provider organization [PPO] vs health maintenance organization
[HMO]), geographic region, and the level of comfort in managing and controlling
utilization. However, both private and government payers share a common process for
collating evidence supporting coverage decisions through the pharmacy and
therapeutics (P&T) committee. This member platform reviews the collective evidence
from drug dossiers, FDA labeling, in-house cost analyses, budget impact modeling,
published treatment guidelines, and physician prescribing patterns to determine
coverage policies (14). P&T committees constitute the organizational decision-making
forum where evidence for drug efficacy, safety and cost are presented, discussed and
evaluated for implementing drug coverage policy. The committees’ reactions to
12

differences in biosimilar attributes can drive the implementation of plan-level policies
affecting the access and potential uptake of biosimilars. Upon market entry, the voting
members of P&T will evaluate product efficacy, safety, and cost evidence to determine
the positioning of biosimilars within the formulary benefit design.
With the expanding contributions of biologic care to future medical practice,
understanding payer preferences in determining the value of biosimilars is critical as
payer decisions affect the evolution of health care delivery. Given the differences
between biologics and small molecule drugs, payers will be evaluating a new set of drug
therapy options under uncertain conditions where differences in clinical efficacy and
safety may have different policy implications. Evidence characterizing product safety,
efficacy, and cost profiles can define these differences and assist in distinguishing
products; however, understanding how payers weigh these factors in formulary decision-
making will provide further insight into achieving access to the biosimilar market.
Eliciting preferences from payers on the attributes most influential to P&T committee
review decisions will provide an objective measurement of payer comfort in managing a
new class of therapies that differs from small molecule generics. Given the novelty of
biosimilars, primary research in discrete choice methods is an appropriate approach for
investigating the decision-making attitudes of key stakeholders, such as commercial
payers, who influence access to these drugs for a significant number of patients in the
US. This research and measurement of payer preferences for managing biologics are
meaningful in providing health service researchers, policy makers and providers an
understanding of how biosimilar market entry in the US will affect health care delivery as
well as patient outcomes.

13

V. Research Objective and Explorative Insights
The proposed objective of this research is to gain insight into the US commercial
payer’s decision-making preferences for novel biosimilar products. Further, this
research aims to understand how uncertainties and trade-offs arising from the FDA
classification of biosimilars and cost compare to innovator products with respect to
formulary management decision-making at various time points of review.
To ground the specific research hypotheses, explorative primary research
interviews (n=17) were conducted in a sample of MCO decision makers. The national
payer sample consisted of pharmacy and medical directors and clinical pharmacists who
contribute to the presentation of evidence at P&T product reviews. Despite the relatively
small sample, the participants were representative of many of the largest MCO plans in
the US which together account for approximately 85 million commercial and managed
Medicaid and Medicare lives.
The primary objective was to probe payer understanding of biosimilar drugs and
identify themes and attributes that influence coverage and managements decisions for
biologic drugs. The focus of the explorative qualitative research was to gauge how
payers will respond to biosimilar market entry and employ utilization management
controls that may vary by drug class and patient population. Additionally, the research
explored how FDA policy regarding the development of an abbreviated pathway may
influence decision-making. The preliminary research aided in identifying the approach
and methodologies most appropriate for investigating both the payers’ reactions and
potential factors influencing the uptake of biosimilars.
Interviews consisted of individual one-hour phone conversations conducted from
January 2011 to March 2011. A discussion guide was implemented to focus the dialog
on addressing the key objectives presented above and included questions intended to
14

establish the payers’ baseline knowledge concerning market entry for biosimilars as well
as the development of an abbreviated pathway.
The explorative research generated key findings, categorized as follows: (1)
background knowledge and current perceptions, (2) assessment of value in proposed
review under uncertainty, (3) expected cost savings with biosimilars compared to
innovator products, and (4) class-specific priorities and utilization management
projections.
Feedback on the payers’ background knowledge and current perceptions
indicated that payers recognize differences between biosimilars and small molecule
generics in terms of molecular structure and manufacturing processes. Although
uncertainty exists around the exact definition of biosimilarity, payers tended to view
biosimilars as alternative branded products rather than small molecule generics.
Specifically, the tendency to define biosimilars as newly branded agents rather than
generics was more pronounced in pharmacy directors than in medical directors. Despite
the limited sample size and inadequate statistical power to test for statistical
significance, differences in clinical and pharmacokinetics/pharmacodynamics training
may have contributed to differences between pharmacy and medical directors with
respect to their understanding of biosimilar structure. In addressing how their
understanding and perceptions of biosimilars may change in the near-term, payers
referenced the FDA guidance and abbreviated pathway development as the major
factors providing evidence thresholds on the safety and efficacy of biosimilars.
During the interviews, payers were queried about how they might assess the
value of biosimiliars under conditions of uncertainty during formulary review. Additionally,
payers were asked to describe the extent to which P&T committee members had begun
considering management strategies. At present, conversations about biosimilar
formulary review and management have been informal and conducted at a high-level.
15

However, payers indicated that contrary to a review for small molecule generics, a P&T
review for biosimilar products will likely vary by class since different drugs and
indications may require different evidence. Supporting these formulary decisions,
clinical pharmacists anticipated that clinical trial data for biosimilars available for the P&T
review will be minimal compared to the evidence supporting innovator drugs. While a
complete Academy of Managed Care Pharmacy (AMCP) dossier will be requested,
payers expect that FDA recommendations, labeling information, and any available
clinical trial data will constitute the crux of the evidence for review.
With biologic drugs representing a sensitive, high-spend area in health care,
payers were also asked about their expectations with respect to the cost savings offered
by biosimilars compared to the innovator/reference products. The majority of MCO
payers anticipated that biosimilars across categories will be introduced with a 10-30%
discount. When probed about cost differentials across various therapeutic classes of
biosimilars, payers indicated that erythropoiesis-stimulating agents (ESAs) and
granulocyte colony-stimulating factor (G-CSFs) biosimilars were expected to offer
slightly higher discounts as these markets are considered more commoditized and the
products are easier to manufacture. Conversely, monoclonal antibodies (mAbs)
biosimilars are expected to enter with lower discounts due to perceived difficulties in
production and the larger expected return on investment with high-priced oncology
drugs. Moreover, payers questioned whether biosimilar mAbs would be available in the
near-term due to manufacturing difficulty and the high burden of evidence potentially
required by the FDA for demonstrating therapeutic similarity.
Payers projected that the utilization management policies for biosimilars will be
contingent on the therapeutic class. Contrary to widespread acceptance of small-
molecule generics independent of drug class, reviews and utilization management
decisions for biosimilars were expected to be situational. Variance in formulary review
16

and placement will largely depend on the drug class budget and the current level of
management within the drug category. These factors influence how healthcare plans
will prioritize and implement utilization management controls. High-spend categories
and drug classes used to treat patient populations with non-fatal chronic illnesses are
most poised for aggressive utilization management. Specifically, payers targeted tumor
necrosis factor-alpha (TNF alpha) as comprising the largest biologic class for biosimilar
uptake due to the high cost of therapy, top budgetary priority, multiple indications with
high patient prevalence, and the high level of physician comfort in disease management.
VI. Refined Research Question and Project Objective
Based on feedback from the managed care payers, it was anticipated that
product value assessments for biosimilars may differ from the case for small molecule
generics and that coverage policies may vary by class. Budgetary and economic
impact, the major drivers in proposed utilization management controls, must be balanced
by evidence demonstrating the relative safety and efficacy of these products.
Importantly, the trade-off of these attributes may be more conservative than for small
molecule generics, which may lead to different management policies and confidence
levels that biosimilars are in “perfect substitutes”. With payers comparing and weighing
the attributes of biosimilar and reference products in determining formulary coverage,
they will be making decisions under greater uncertainty and more limited data than that
of generic substitutes.
Applying the literature on current formulary decision-making, discrete choice
methodology, and the payer insights acquired through explorative qualitative interviews,
the principal aim of this project is to elicit US commercial payer preferences in evaluating
the value of biosimilars, and to gain insight into how payers prioritize product attributes in
making management choices for formulary placement and market access. To conduct
this research objectively and yield quantifiable results, discrete choice methods will be
17

used to investigate payer preferences. The value of this research and modelling is to
capture how attributes of biosimilar drugs are weighed by payer decision-makers and
ultimately how these preferences drive access policies for US managed care
organizations.

18

Chapter 2: Literature Review
This chapter will review the relevant literature on the use of choice behavior
methods including conjoint analysis and discrete choice modeling, as well as the
potential model specifications and estimations used in healthcare decision-making. In
addition to decision analysis and stated preference (SP) methods, the chapter will
highlight the use of these methods in the other disciplines and contexts, specifically in
the areas of management and organizational business, and discuss how these methods
can be leveraged and applied in healthcare decision-making.
I. Background on Choice Behavior
Methods used in examining choice behavior are rooted in traditional
microeconomic theory and consumer behavior theory wherein preference model
specification is characterized by a utility function derived from the properties or attributes
of a particular good (15, 16). Despite the inability to account for all attributes and
characteristics that may contribute to choice decisions, specifying a random utility model
(RUM) allows an analyst to explain preferences with the utility function by assuming that
the differences in the unobserved variables of preference alternatives are related to the
choice probability distribution within the population (17). In order to operationalize and
empirically identify this choice model, a distribution function of choice probability must be
specified. Much research on discrete choice modeling has utilized the extreme value
type I (EV1) distribution leading to the basic multinomial logit model (MNL). McFadden
and others have demonstrated that with a key set of axioms and assumptions and
utilization of an EV1 probability distribution, one can translate the unobserved random
component of each alternative into a definable component of the probability expression.
This leads to the MNL choice model where the unknowns are utility parameters
associated with each observed attribute of the utility expression through convenient
maximum likelihood estimation methods. The widespread use of this basic model has
19

been explored and expanded to address potential violations of the independence-from-
irrelevant alternatives (IIA) assumption, and to allow for heterogeneity in taste weightings
among the population. Chapter 3 discusses in greater detail potential expansions of this
model and specific application for use in study design.

II. Application of Discrete Choice Methods in Healthcare Valuation
The application of choice modeling methods in health care has the potential to
explain the impact of various drivers that influence the levels of health care decision-
making for a treatment or an intervention. Specifically in the case prior to market
approval or availability of treatment where true consumption behavior cannot be
captured, choice modeling and conjoint analysis are effective tools for understanding
stakeholder and patient stated preferences. The various levels of healthcare delivery in
the US and other regulated markets comprise of multiple stakeholder levels where
decision-making influences treatment selection. These involve regulatory bodies,
managed care and payer access assessments, physician prescribing preferences, and
patient behaviors and value perceptions.
Many researchers have recognized the potential benefits of using discrete choice
modeling in health care to gain insight and to elicit stakeholder preferences, particularly
with respect to how decision makers trade off attributes of healthcare intervention
options (18-21). To elucidate the value of this methodology in healthcare decision-
making and explore valid applications, the International Society of Pharmacoeconomics
and Outcomes Research (ISPOR) formed a Conjoint Analysis Good Research Practices
Task Force to develop criteria for good research practices governing discrete choice and
conjoint analysis (22). With identifying guidance on good research practices, the Task
Force highlighted preferences and valuations that have been previously studied and
measured across a broad spectrum of healthcare interventions, including: cancer
20

treatment (23,24); HIV prevention (25), testing (26), and treatment (27); dermatology
services (28); asthma medications (29); genetic counseling (30,31); weight-loss
programs (32); diabetes treatment (33) and prevention (34); colorectal cancer screening
(35,36); depression (37) and treatment for Alzheimer’s disease (38).

III. Discrete Choice Methods in Healthcare Decision-Making:
In addition to applying discrete choice methods and conjoint analysis to evaluate
preferences concerning the different healthcare interventions as listed above, these
methods have been more recently applied with the aim of understanding decision
makers’ preferences for newer therapies that are more effective but associated with
therapeutic risks and the presence of uncertainties (38). Specifically, the application of
these methods may address the many recent innovations in biologic medicines for which
medical decision makers often deliberate over safety and efficacy trade-offs in selecting
treatments. Although much of this research has been centered on patient preferences
for novel treatments, conjoint analysis methods have been also employed to understand
physician prescribing and clinical decisions and to gain insight into how different
stakeholders might value health outcomes (39).
In conjunction with research examining physician and patient preference
decision- making in healthcare, it is important to recognize the value of decision-making
behaviors in other stakeholders contributing to the access environment in which
physicians and patients are able to make individual treatment choices. Specifically, this
research focuses on organizational decision- making at the regulatory and
reimbursement levels which largely influence market access for particular treatments.
Exploring research at the healthcare delivery systems level expands choice behavior to
include policy formation and risk management at the population level.
21

Literature examining regulatory and managed care decision-making have linked
the downstream effects of these policy decisions to outcomes, including drug approval
and labeling (40, 41), health outcomes (42), therapeutic treatment choices and
prescribing behavior (43), utilization management (8, 44, 45, 46), and healthcare cost
mitigation (5, 47, 48). Although the breadth of this research provides insight into how
these access policy decisions relate to health outcomes, value, and the costs of health
care, the evidence does not address the factors contributing to the decision trade-offs
faced by stakeholders in formulating policy.
Payer and regulatory decision-making processes have attracted the interest of
manufacturers who invest heavily in the development of medical technologies and
therapies while anticipating market approval and reimbursement. Accordingly,
pharmaceutical and biotechnology manufacturers are particularly focused on generating
evidence to demonstrate the value of products to decision makers who will ultimately
regulate and fund these drugs and interventions. To understand these preferences and
how they may change throughout the drug lifecycle, vendors conduct qualitative
research on managed care behaviors and regulatory decision-making through expert
panel advisory boards, primary research interviews, and focus groups. Manufacturers
have the advantages of utilizing these methods to understand decision-making behavior
by targeting discussions with key stakeholders and testing value messages to
understand how they may influence future decision making. However, in the interest of
external investigators and researchers, the availability of the implications and results of
this research are limited. Qualitative methodological approaches are not clearly rooted
in economic consumer theory for decision-making and therefore do not allow for
systematic trade-offs to be evaluated and measured. Highly proprietary, the majority of
this research is restricted for use in internal marketing and promotional strategies and
therefore, is not available in the public domain.
22

a. Opportunities for Expanding Research
With the expansion of conjoint analysis and discrete choice methods outside of
societal medical intervention valuations and patient preferences, work by FR Johnson
and colleagues has contributed to the understanding of health technology adoption and
reimbursement decisions (49). Specifically, a two-stage conjoint analysis experiment
was used to elicit preferences for technology adoption criteria and thresholds for
reimbursement recommendations from a sample of ISPOR participants. The study
focused on eliciting reimbursement recommendations and product valuation preferences
from a European perspective by including attributes of the National Institute of
Excellence (NICE) adoption criteria. Overall, the study demonstrated the feasibility of
using SP methods and decision modeling as a means for quantifying preferences in
organizational-level stakeholders. Additionally, the survey identified pharmaceutical
drug attributes potentially relevant to a policymaker as well as the associated trade-offs
involved in deciding product preference.
As previously mentioned, the decisions of US payers concerning biologics and
specialty drugs largely influence physician preference and patient use. Payers can
potentially impose heavy restrictions on utilization and cost-sharing controls that affect
treatment choices. With the market approval and entry of biosimilars in the near-term,
sensitivity to cost and risk is becoming critical to decision-makers who are evaluating
product value under new legislation and regulatory adoption criteria. Qualitative and
anecdotal evidence from manufacturing research can help guide understanding of how
US managed care commercial payers will make access decisions for biosimilars;
however, more sophisticated approaches for exploration are needed to minimize the
subjectivity associated with qualitative research. Extending the work in organizational
decision-making using discrete choice modeling and conjoint analysis methods, the
23

following chapter describes how preferences for managing novel biologics can be
systematically examined.

IV. Use of Organizational Decision Making Methods in other Economic Sectors
Despite the scarcity of quantitative organizational decision-making research in
healthcare, methods for decision analysis have been implemented across many other
disciplines where researchers wish to elicit organizational decision-maker preferences
under uncertainty. Within this context, researchers derive insight by evaluating potential
alternatives and applying decision analysis methods to calculate the value of additional
information not yet available to the decision-maker. Although broadly applied, the
common objective of decision analyses is to determine the optimal option for
organizational decision makers which is consistent with the given information, and to
understand how the attributes of alternative options can potentially impact the overall
value of the selected choice. Evidence and results across disciplines may reveal that
the values for alternatives are not aligned with the intended organizational perspectives.
This highlights an important limitation of these methods since this approach is not
intended to solve a decision problem directly (50).
Notable decision analyses have been conducted in the research and
development of new products and the adoption of policies within both the private and
government sectors. Specifically, as part of the U.S. energy policy considerations, a
decision analysis framework was used to analyze the value of uranium resource
information (51). Similar to the unknowns of how information on biosimilars may develop
and change preferences for value over time, the research demonstrated the application
of decision analysis methods for determining how decision makers value unknown
information and showed the impact of discovering this information on changes in
preferences over time. Other government branches and public services have applied
24

decision analysis for other purposes, including efforts to understand the possibility of
seeding hurricanes threatening the coasts of the U.S. (52), protection from wildfires (53),
and various strategies to promote school integration (54). Other exemplary uses in
developing fuel emission control strategies (55) and managing chronic oil discharge (56)
illustrate how these methods can aid in formulating policymaking standards.
As noted previously, much of the decision analysis research conducted in the
corporate sector is largely proprietary. However, some work focusing on organizational
decision-making by corporate boards and the policies created to allocate project
investment and company budgetary considerations have also been cited (57, 58, and
59). These studies also demonstrate how SP methods are used to gain decision-making
perspective and establish decision rules that set organizational expectations and
policies.
V. Organizational Decision Making Perspectives in Policy Adoption
A specific subset of decision analysis research focuses on policy-implementation
methods. Researchers frame decision constructs to elicit preferences on policy structure
and formation using alternatives with given information sets. An overview of this
literature is presented herein to illustrate how organizational decision-making bodies
have employed decision analysis methods to formulate downstream policies affecting
internal structure.
To create a valid context for decision-making at the organizational level,
representative situations and alternatives must be presented and accurately depict the
realism of the true decision and role of the subject of interest. Along with selecting
appropriate alternatives and attributes to frame appropriate decision contexts, the range
and frequency effects for attribute levels have also been shown to influence attribute
weighting and quantitative preference outputs (60). Although capturing all significant
information in light of imperfect knowledge and cognitive capacities is challenging, there
25

is evidence that individuals tend to base judgments on a relatively small number of
criteria (61, 62, and 63). As noted in the literature, one useful approach for capturing all
relevant information for experimental design is to conduct an explorative pilot survey of
subjects similar to the decision-making sample (60). Examples of this method have
been presented in the management literature for evaluating job choice, candidate
selection criteria, and compensation policies (64, 65, and 66). According to the
researchers, surveying decision makers themselves to develop the constructs for an
experimental design a priori can enhance the real-world credibility of the survey and
policy implications.
Related to decision-making and policy research, discrete choice and conjoint
analysis have been applied and cited extensively in market research as a means for
eliciting consumer preferences for products and services. In collecting decision-weights
and outputs, conjoint analysis extensions have contributed to the understanding of
organizational decision-making and policy formation constructs.
Discrete choice methods have been applied in organizational decision-making in
an effort to understand how venture capitalist groups determine investment
opportunities. Judgment and preference weighting are a fundamental part in making
actionable decisions in the entrepreneurship process (67, 68, and 69).
Central to the entrepreneurship process is the concept of understanding how
decision-makers select choices under a certain level of risk and uncertainty, which
parallels payer perspectives in healthcare evaluation. Studies on decision-making
processes involving risk perception have identified trends and themes that support the
use of these methods in studying MCO decisions for formulary coverage.
Studies on risk management and the measurement of risk attitudes in business
stakeholders have examined the influence of experience level and individual taste on
investment decisions. Notably, studies comparing entrepreneurs to non-entrepreneurs
26

found that the two groups categorized risk attitudes differently and had significantly
different subjective interpretations of high-risk choices compared to other business
stakeholders (70, 71, and 72). This evidence suggests that hierarchal roles and
perspectives in decision-making may facilitate different business decisions depending on
the role of the decision-maker. Analogous to the structure in entrepreneurship and
management decision-making, healthcare delivery decisions include both organizational
perspectives from managed care payers and prescribing decisions from physicians.
Payers’ perspectives are driven by their role to manage risk across populations whereas
physicians’ decision-making is characterized by their role to prescribe treatments to
individual patients. Research using similar discrete choice methods may offer insight into
how payer and physicians differ in their attitudes toward risk in making decisions for
managing novel therapies.
Subsequent to weighing opportunities and evaluating their associated risk, the
literature on organizational business decision-making also explores how policies are
implemented to maximize the net benefit to the organizational body. Creating new
opportunities involves selecting the most valued decision option and creating an internal
policy to support the decision by coordinating the multiple roles and functions within the
organization. Conjoint analysis and discrete choice methods have been used to
examine the development of internal processes and policies that establish trust among
various levels of stakeholders within the organization (73, 74, and 75). In the extension
to healthcare, managed care decision makers must consider how their formulary
decision policies will affect physicians in their role of prescribing treatments to patients.

27

Chapter 3: Study Design and Methods
This chapter describes the proposed research design approach and methodology
including survey development, the stated hypotheses, the US payer sample, recruitment,
and the model estimations. Additionally, potential limitations of the study design, model,
and sample are discussed.
In order to understand the cost mitigation opportunities that market entry of
biosimilars can potentially yield in the US, it is necessary to investigate the perceptions
of managed care payers concerning product value (relative to the originator product) as
well as their preferences for formulary management. Discrete choice modeling provides
an appropriate, attractive vehicle for eliciting these preferences and more importantly, for
identifying the factors that influence selection. By utilizing discrete choice exercises, this
study aimed to measure stated therapy preferences for the proposed biosimilar product
profiles and to examine the hypothetical formulary management policies potentially
implemented by payers to drive market access.
I. Survey development
a. Product Attributes and Proposed Levels: Selecting appropriate product profiles
requires that the attributes exhibit external validity in their importance to the survey
sample, are consistent with identified factors in medical decision making, and follow the
appropriate model assumptions used to estimate the utility weights.

i. Therapeutic class: With many biologic drugs available and extensively
utilized in the US, the primary concern in selecting appropriate
therapeutic classes for this exercise was to ensure that these drugs
represented therapeutic areas for which biosimilars are highly anticipated
and likely to enter the market in the near-term. Along with contributing to
the real-world applicability of the study, the proposed drug classes
28

represent high-interest areas for which cost-mitigation and management
control are concerning to US payers. The proposed drugs categories
include: anti-tumor necrosis factors (TNF), oncologic monoclonal
antibodies (mAb), and human growth hormone (HGH). TNF therapy
represents a high-budget area for payers; spending in this specialty
category has been reported to exceed all other specialty disease states.
In 2008, TNF therapies accounted for 27.5% of the total specialty
pharmacy spend in 2008 (Specialty Drug trend Report, Curascript 2008).
The development and introduction of biosimilars in this drug category is
also underway in global markets where regulatory barriers and patent
protection rights do not limit entry for this drug class. With the opportunity
for cost-savings in the background of chronic therapy and high disease
prevalence, inflammatory conditions including rheumatoid arthritis,
Crohn’s disease, ulcerative colitis, and plaque psoriasis represent
relatively low-risk patient populations in which substitution with a novel
biosimilar is more attractive to payers and healthcare providers. HGH
therapy was also proposed because it is a relatively commoditized
biologic class with multiple competitors. In the previous qualitative
explorative research, the US payers perceived this drug class as being
low-risk for product substitution. Additionally, HGH follow-on entrants are
regarded as competitors managed by a regulatory framework closely
resembling that for future biosimilars. Payers perceive that the guidelines
for abbreviated approval and the required clinical evidence for biosimilars
will closely resemble that for HGH. In the US, the HGH drug class serves
as a benchmark and proxy for current biosimilar competition and
management. Although the FDA approval of these drugs falls under an
29

alternative process (section 505 b Federal Food, Drug, and Cosmetic
Act), many payers believe that HGH competitors exemplify biosimilars
prior to the abbreviated pathway development. Oncologic mAbs
represent a high-expenditure specialty category that is highly sensitive for
both payer management and physician comfort. Physician preferences
and cancer treatment guidelines are reflective of the sensitivity of cancer
patient populations. Therefore, potential for payer management controls
in this category may be limited.
The selected three drug classes serve as distinct categories wherein
differences in payer perceptions are expected. Furthermore, differences in
biosimilar/originator preference across drug classes are hypothesized to
affect payer management attitudes.
ii. FDA Classification: Product safety and efficacy profiles are highly
influential to payers and prescribers in generating treatment algorithms
and treatment selection, respectively. Although payer perceptions on
clinical endpoints and adverse event profiles are examined closely during
the P&T committee review, payer decision-making does not include the
evaluation of sufficient safety and efficacy outcomes for market approval.
Market entry and approved use are determined by the appropriate
regulatory pathways instituted by the FDA. In February 2012, the FDA
published three draft guidance documents outlining the safety, efficacy,
and quality measures proposed for use in evaluating biosimilars and
interchangeable drugs for market approval. Although the guidelines
provide a regulatory framework for abbreviated approval, the specific
safety and efficacy evidence requirements may depend primarily upon the
specific therapeutic class, and must therefore be evaluated on a case by
30

case basis. Regulatory approval of these drugs includes a thorough
evaluation of product safety and efficacy criteria whereby higher
thresholds exist for interchangeable compared to biosimilar drugs.
Therefore, payer decision-making and the acceptance of biosimilars will
correlate strongly with FDA guidance on how these molecules are
classified using the abbreviated pathway criteria. In the discrete choice
survey, innovator, biosimilar, and interchangeable classifications will
comprise the levels as these serve as constructs capturing the latent
measurement of safety and efficacy perceptions of payers.
iii. Cost: Cost of therapy, a factor to which both payers and patients are
highly sensitive, contributes largely to formulary placement decisions and
the associated and cost-sharing levels. With contracting and rebate
incentives offered by manufacturers, the relative costs to specific
managed care health plans may differ. Estimated price differences
offered by biosimilars have been modeled using the price ratios of
branded small molecules to generics and accounting for the higher fixed
costs for developing and manufacturing more complex biologic molecules
(10,11,12). Drawing from these projections, price levels will be derived as
percentage differences between biosimilar and originator products. Using
a net price difference rather than absolute costs in US dollars is expected
to elicit payer reactions and considerations for product cost including
potential contracting and rebate levels. Based on supportive data
modeled and projected from the literature, cost differentials between
innovator and biosimilar/interchangeable products will range from <10%
to > 30%.
31

iv. Time: Upon product launch, timing of the payer P&T investigation and
drug preview may vary depending on the timing of product entry as well
as the existence of other high-priority products that require abbreviated
review for immediate clinical access. Although timelines for drug review
and policy publishing requirements vary by managed care plan, formulary
decisions at launch are limited to clinical trial evidence with available ex-
US post launch data and are considered temporary until a full-review is
completed; this occurs approximately 6-12 months post-launch. With
limited real-world data on effectiveness and in the absence of plan-level
utilization data at the time of product market entry, decisions at launch are
intended to provide drug access. In the interim, further evidence and data
on physician prescribing behaviors and patient outcomes can be collected
for a more extensive review. Following a full review and the
implementation of published policy on formulary placement, payers may
conduct subsequent class reviews periodically in response to clinical
guideline changes or lapsed contracts where there is opportunity for
renewing discounts to manufacturers. To account for the preferences
differences related to timing of the product launch and to closely
approximate the time intervals in which payers would most likely review,
the levels for time are defined as launch, near-term (6-12 months), and
long-term.

b. Sequence of Experiments: The survey will consist of a two-stage discrete choice
experiment in which the first question relates to product preferences (i.e.
biosimilar/interchangeable or branded innovator) and the second follow-up question
relates to the formulary management decision (i.e. parity access). Pairs of product
32

profiles with the attributes listed above will be presented to survey participant payers
who will be asked to select which product they believe is superior. To further investigate
the strength of product preference and the payer attitudes regarding class management,
the second question will ask the payers to select the formulary positioning of the superior
agent. Managed care plans may operate under various levels of management control in
their formulary design; however, stronger preferences for preferred drugs are
incentivized through fewer restrictions and better formulary access. To account for the
ordinal spectrum of formulary management controls potentially implemented by payers,
the second discrete choice question will present the following four response buckets: (1)
biosimilar not covered, (2) biosimilar parity access, (3) biosimilar preferred (no patient
switching), and (4) biosimilar preferred (patient switching). These response buckets
represent varying degrees of access which payers can achieve through different
utilization management techniques (ie, tier differential, step-therapy through the
preferred agent, National Drug Code [NDC] block). State variation in healthcare
coverage legislation may prevent particular payers in certain US regions from using
certain management techniques, such as mandates to switch existing patients; however,
this option is included in the present research to capture the most aggressive
management option employed by many HMO’s and closed-model systems.

c. Profile Generation: In order to capture the main effects and two-way interaction terms
among attributes, a fractional factorial method will be used to select profiles for
comparison.

d. Attribute Ranking Exercise: An attribute ranking exercise will be included in the
survey where payers will be asked to rank the importance of each of the product
attributes in descending order. Although not part of the primary research objective, this
33

explicitly-ordered ranking will be compared to the implicitly-derived preference weights
from the choice model.

II. Hypotheses:
The proposed objective of this study is to understand payer access and formulary
decision-making for biosimilars upon market entry. Using discrete choice exercises with
US commercial payers, stated preferences for biosimilar product attributes will be
measured and compared for therapeutic class effects. Furthermore, preferences for
utilization management controls for biosimilar products will be measured and compared
across categories to reveal payer comfort and willingness to manage across therapeutic
categories. Using qualitative insight gained from the explorative interviews and previous
literature on formulary management controls in specialty disease areas, the following
primary hypotheses are identified:
a. With larger price differentials offered between biosimilar and branded
biologics, payers will prefer biosimilar products (effect of price on product
preference)
b. With larger price differentials offered between biosimilar and branded
biologics, payers will prefer more aggressive utilization management policies
(effect of price on access management)
c. Payers will be less aggressive in utilization management for biosimilar drugs
in oncology than in the TNF and HGH categories.

In addition to hypotheses around attribute effects, one additional
secondary hypothesis is defined to explore explicit vs implicit elicitation of
biosimilar attribute ranking (d).

34

d. Explicit vs implicit attribute weighting: Using the output from the attribute
ranking exercise, explicit attribute and implicit rankings will be compared for
consistency and priority. Specifically, payers may be more implicitly sensitive
to price (as revealed by the utility estimates from the choice model) than the
explicit ranking given when compared to other attributes.

III. Sample
US Payer Sample: The survey sample will be obtained from the population of
US managed care payers. Collaboration with a pharmaceutical value strategy
consulting firm allows access to a database of payers providing a convenient
sample pool. Drawing from this convenience sample of payers actively involved
in payer research with various consulting projects maximizes the survey
response rate compared to other methods for recruitment. Although this
database is not exhaustive of all US payers representing all managed care plans,
the database includes national and regional payer organizations representing
approximately 85 million covered commercial and Medicare lives. To ensure a
valid sample that accurately represents payers who contribute to formulary
decision making and positioning, potential participants are screened using
inclusion criteria. The criteria specify that participants are voting members of the
P&T committee and have an identifiable role (i.e. pharmacy director, medical
director, clinical pharmacist) in the organization. Participants must verify that they
have been in their present role for at least two years in order to ensure their
familiarity and experience with the P&T process. Additional questions to gauge
the payers’ background knowledge of biosimilars and the development of the
abbreviated pathway legislation will be posed to certify that the participants have
35

a baseline understanding of biosimilar drug classes and can meet the potential
timeline for availability.
Sample participants will be recruited by e-mail providing information on
the research objective, discrete choice exercises, and disclosure. To ensure
adequate sample size, target recruitment from the database is estimated to be
approximately 25 payers with each payer completing approximately 11
experiments. A balance between pharmacy and medical directors is preferred to
provide a representative and broad view of clinical expertise and also to test the
attribute ranking secondary hypothesis.

IV. Model Specification and Estimation:
Model Selection: Choice models are used to understand consumer
preferences and utility weights for meaningful attributes that drive decision
making. The most commonly used models that allow for the estimation of utility
weight parameters are the multinomial logit (MNL) and conditional logit (CL)
models, which specify for individual and attribute specific covariates,
respectively, that contribute to the choice probability.
Although these models have been used within many different contexts in
decision-making research and have convenient closed forms for estimation, there
are several restrictions and limitations to using these simplified models that may
violate the choice experiment data generating process described here.
Specifically, these models require an assumption of homogeneity in attribute
weights across the population. This restriction implies that all US payers weigh
attributes in the same way and will exhibit utilization management decision
behavior similarly. Given the US payer sample will draw from national and
regional managed care organizations operating in a variety of different health
36

care delivery models (eg, PPO, HMO, fee-for-service [FFS], open formulary
system, etc), P&T committees may employ different management ideologies
when deciding how new products will be incorporated into the benefit. These
differences in plan characteristics and the respective levels of management
control may imply that utility weights vary by plan as well as by the individual P&T
committee member. A more flexible model accounting for heterogeneity in
attribute weighting relaxes the restriction of uniform attribute weights across the
population. To achieve this, the random parameters logit (RPL) model is
considered.
Relaxing the homogeneity assumption, the RPL model allows for a
unique attribute weight to be estimated for each payer rather than restricting
payers to value each attribute and management control in the same way. The
model assumes that attribute weights are distributed amongst payers with a
known statistical distribution and then identifies a specific deviation for each
individual from this average weight. The flexibility of this specification is
appropriate considering that the sample will include US payers from different
managed care organizations that may have different taste weights for attributes
and management controls.
In addition to the advantage of relaxing the homogeneity restriction, RPL
is not bound to the independence or irrelevant alternatives assumption (IIA) as in
the case for both MNL and CL models. RPL allows individual taste weightings to
deviate from the average sample taste weighting, although this deviation is not
directly observed and therefore enters into the unobserved portion of utility in the
equation. In turn, the unobserved portion of utility is correlated across choices, as
the individual’s deviation is the same for each choice, and thereby avoids the
problem posed by the IIA.
37

In considering a more flexible model like the RPL (in contrast to the MNL
or CL model that may target specific considerations for the proposed
experiment), it is important to address the feasibility for estimation given that the
RPL model cannot generally be estimated using standard maximum likelihood
techniques since this integral does not have a closed form. Simulated maximum
likelihood is the commonly used method to estimate RPL. It presents additional
costs in estimation where for the given average sample taste weight parameters,
an individual taste weight is drawn from a distribution and used to calculate the
conditional logit probability. This process is repeated many times and the
average conditional logit probability is considered the approximate choice
probability. Despite the advantages afforded by the heterogeneity of payer
preference weightings, this model requires a more complex computing platform
for compared to the CL or MNL, rendering this approach burdensome.
Random Utility Theory (RUT) modeling as an additive specification:
To consider the use of a random utility model for choice behavior and
specification and estimation of the RPL model for the proposed experiment, it is
important to consider potential violations that may exist using an additive model.
Particularly, the zero condition assumption can potentially be violated if for a
fixed level of an attribute (ie, time), no trade-offs are made and the probability of
choosing an alternative does not change. In this case, the model would estimate
utility parameters that are not consistent with payer preferences by
overcompensating the utility parameters for the other attributes in the profile. A
case where violation of additive preference structure may have occurred was
uncovered in the pilot exploratory research for oncologic mAbs. In this particular
therapeutic class, payer preferences for management and biosimilar adoption
were largely inhibited by physician discretion in cancer treatment choices.
38

Payers are reluctant to manage oncology products and defer largely to physician
choice to guide uptake, particularly at launch of new therapies. To account for
this behavioral assumption, it may be appropriate to omit the oncology mAb
biosimilar comparisons at launch (ie, time zero). In application, excluding this
profile option from the experiment acknowledges that in the case of this class of
biosimilars, managed care payer preferences for product attributes will not be
additive. To test this assumption, a secondary experiment should address this
potential preference structure at these levels to determine if perhaps a
multiplicative model structure might be more appropriate to measure preferences
for biosimilar in this class at launch.

39

Chapter 4: Results
The chapter presents the results of the primary research with US managed care
payers including both the estimation of the quantitative model as well as the qualitative
discussion findings from the survey and interview. Both of these findings are useful in
assessing the sample’s preferences regarding biosimilars. For the model estimation, the
attribute weighting and marginal weights are provided. Qualitative results are organized
and summarized according to four key themes that emerged during the interviews.
I. Survey Sample
Recruitment e-mails were sent to prospective participants between January and
March of 2013. Payers were asked if they would be willing and interested in
participating in a one-hour phone call at their convenience to discuss their perceptions
and perspectives on biosimilar adoption in the US. A total of 61 emails were sent;
approximately half of the recipients responded and agreed to participate (n=30).
Twenty-eight interviews were confirmed and successfully scheduled between January
and the end of March. Due to scheduling conflicts, the remaining two respondents were
not confirmed. Table 1 presents a descriptive summary of the survey sample including
plan representation (national vs. regional) as well as managed care role (pharmacy
director vs. medical director). At the time of the research fielding, the sample payers
represented the following managed care plans: Aetna, Blue Cross Blue Shield, Cigna,
Coventry, Excellus Blue Cross Blue Shield, Gessinger, Harvard Pilgrim, Health First,
Health Net, Health Plan of Nevada, Humana, Kaiser, Amerihealth Mercy, MedVentive,
Oxford Health Plan, Presbyterian Health Plan, Select Health, United, WellCare Health
Plans, and WellPoint.

40

Table 1: US Managed Care Payer Sample Description
Variable Total
Interviews
Medical
Directors
Pharmacy
Directors
National
Plans
Regional
Plans
N 28 12 16 15 13

Each respondent completed approximately 11 comparisons although some variation
existed depending on the time during the interview and how quickly each exercise was
completed. In total, 300 observations were collected for analysis.
II. Qualitative Results
To provide context for survey responses and capture qualitative information
about payer awareness on biosimilars, open-ended questions were asked in addition to
the discrete choice exercises. Qualitative questions focused on several key topics: (1)
the definition and understanding of biosimilars, (2) awareness of biosimilar regulatory
pathway development, (3) decision-making process and review for biosimilars by the
plan, and (4) future benefit structure changes that may support biosimilar management.
Several key themes emerged during the discussion, as described below.
a. Biosimilars compared to small molecule generics
The payers reported a high level of understanding of how biosimilars differ from
small molecule generics. Subtleties in biologic structure and production were
recognized as the main reasons underlying these differences. Payers agreed that these
differences may be associated with variance in safety and efficacy between molecules.
However, a majority of payers stated that these variances also exist within batches of
the same reference molecule (ie, “product drift”) and this concern is not unique to
biosimilar copies only. In summary, payers believe that the uncertainty in clinical effect
from biosimilar use closely resembles the uncertainty and risk harbored by branded
biologics at market approval.
41

Payers also viewed the manufacturing processes for biologics to be significantly
more complex than for small molecule drugs, noting that the regulation of biosimilars
primarily depends upon the quality and consistency of the production process as well as
final product. US payers indicated that quality control guidance and regulation by the
FDA for manufacturer’s novel to the biotechnology manufacturing market may need to
require data demonstrating both product safety and manufacturing consistency for
quality. One concern for newly establish biotechnology manufacturers is the ability to
ensure consistent supply and distribution of product as a part of contract negotiations.
b. Awareness and effect of FDA Regulation
During the course of the present research, the FDA published several guidance
documents on the abbreviated regulatory pathway. Although not confirming or finalizing
pathway guidelines, these guidance documents informed how the FDA differentiates
data requirements and review processes for market approval for biosimilars from those
for small molecule generics. Particularly, a key theme in determining biosimilarity is the
use of a “step-wise approach” in data generation and regulatory evaluation. Through
this process, the uncertainty about the biosimilarity of the proposed product is identified
and informs the steps and evidence required to address this uncertainty. This implies
that biosimilar market approval and data submissions will be specified case-by-case
where data requirements for each abbreviated biologic license application (ABLA) may
vary by product. Additionally, the FDA notes that the overall approach in generating
these data thresholds for biosimilar market approval will be based on the “totality of
evidence,” reinforcing that regulatory decisions may vary by product and availability of
analytical methods to demonstrate biosimilarity at time of submission. Two important
themes emerge from the FDA guidance documents which are relevant to the current
research and relate to (1) how payers perceive the FDA requirements for biosimilars,
and (2) how they are differentiated from small molecule generics. The focus of the
42

qualitative questions around payer awareness of the FDA review is to understand how
regulatory considerations, evidence, and product labels inform how payers evaluate and
manage biosimilars.
US payer awareness of the FDA pathway development was moderately high, but
knowledge about the specific data requirements and recommendations was not
universal. Payer familiarity with the pathway guidance varied; the majority of payers
agreed that until a pathway was finalized, the guidance bears a relatively small impact
on the decision-making process of the MCO. The consensus is that by specifying
pathway requirements, the FDA is responsible for determining and evaluating safety and
efficacy criteria for biosimilars and ensuring they are adequately comparable to that of
the reference products. Payers acknowledged that the FDA criteria may differ by the
class and complexity of the biologic molecules. Ultimately, payers will yield to the FDA
to ensure biosimilars are safe and efficacious. Payers also noted that biosimilar quality
considerations and production processes are important issues that the FDA will likely
monitor in a similar way that reference biologics are monitored.
During the qualitative research, payers were probed on their understanding of the
FDA designations of biosimilarity and interchangeability in order to determine how they
perceive the clinical value of an interchangeable product. Related to this, another
discussion focused on how payers may manage interchangeable products differently
than drugs approved as a biosimilar.
The majority of payers were uncertain about how the FDA will assess
interchangeability and whether this designation will allow for automatic substitution.
Additionally, payers were interested in how coding for biosimilars and interchangeable
products might differ. Without a product being AB-rated as in the case for small
molecule generics, payers may be unable to enforce automatic substitution for
interchangeable products. In this case, payers indicated that discerning the value of
43

interchangeable products may be directed to how physicians perceive their use and
willingness to substitute therapies in their patients. Without the authority and regulation
of these products to be substituted at the pharmacy level, physicians may be the key
decision-makers in determining the uptake of interchangeable products. The payers’
willingness to aggressively manage interchangeable products that are not AB-rated
appeared to be guarded where forced substitution is unlikely, especially for those
patients who are currently on stable therapy with a reference brand. Payers noted that
although physicians may be key decision makers with adoption, their formulary
management decisions will incentivize the promotion and utilization of biosimilar and
interchangeable products. These findings will be discussed in greater detail in the
section outlining the changes in future benefit structure considered by payers.
c. Formulary review and decision-making process for biosimilars
With respect to the evaluation of biosimilars and formulary review, payers noted
that current discussions are conducted informally and at a high level. At the time of the
present research, a biosimilar had not yet been approved on the market to review, and
therefore no formalized P&T policy discussions had occurred. When asked to describe
the formulary review process for biosimilars, payers indicated that the same key
attributes (safety, efficacy, and cost) would be evaluated to support any changes in
coverage. This information supports the notion that payers will be using the same
process and criteria for biosimilar evaluation. Despite the previously determined,
nonspecific responses regarding decision-making criteria, the majority of payers cited
Omnitrope (human growth hormone) and Extavia (interferon beta 1-b) as two examples
of biologic molecules that they consider analogous to biosimilars with respect to
formulary review.
Omnitrope, somatropin, a recombinant HGH product substantially similar to
Pfizer's Genotropin, was approved in 2006 under section 505(b)(2) of the Federal Food,
44

Drug, and Cosmetic Act (FDC Act). Section 505(b)(2) provides an abbreviated approval
process for pharmaceuticals, including follow-on protein products that are significantly
similar to other protein products approved under the FDC Act. Although the majority of
proteins and biologics approved in the US do not fall under this pathway, payers note
that the arrival of Omnitrope incited the FDA to develop an abbreviated biologics
approval process for more complex protein and biological products. The majority of
payers indicated that formulary review of Omnitrope closely resembled the review of
other biologic products in therapeutic categories where multiple branded agents are
available. Specifically, this molecule was considered to be a relatively simple biologic
molecule where the existence of multiple brands and mature market availability
facilitated a streamlined formulary decision-process. Payers believe that with the advent
of biosimilars in therapeutic classes that exhibit more molecular complexity, the
evaluation process will be comparatively more rigorous than that for Omnitrope.
Specifically, attention to the data and evidence for product safety will be critically
important.
Extavia, interferon beta 1-b, was approved in 2009 for the treatment of relapsing
forms of multiple sclerosis to reduce the frequency of clinical exacerbations. Approval
for Extavia was achieved through a BLA (biologics license application), the regulatory
approval pathway for the majority of branded biologics in the US. Payers observed that
the approval of Extavia and the product label closely resembles that of another interferon
beta 1-b, Betaseron. Approximately half of the payers surveyed indicated that this
analog serves as a predecessor for the potential process for biosimilar review. Payers
noted that data on safety and efficacy are paramount in determining coverage.
Additionally, product price relative to all others in the category will determine the level of
access. In particular, two plans indicated that with Extavia available at a lower cost
45

compared to Betaseron, all new patients are required to use Extavia first unless a
physician appealed to use Betaseron.
With regards to formulary review, payers were posed an open ended-question
about their expectations for the pricing of biosimilars. All payers indicated that
biosimilars will cost less than the innovator biologic but noted that price differentials for
biosimilars will not resemblethe price differences for small molecule generics. Payers
projected that biosimilars will cost approximately 10-30% less than innovator brands.
Factors potentially facilitating a more precise range for these prices include the number
of available biosimilars in the market, interchangeability status, and the current level of
rebating/contracting existing within the therapeutic class. Contrary to the approach
payers take with small molecule generics where decisions are consistent for the majority
of all drugs, the biosimilars therapeutic class may influence how payers prefer and place
them into the formulary.
d. Future Benefit Designs
In order to frame the discrete choice experiment with respect to the selection of
formulary management decisions, payers were posed an open ended question about
how they envision using biosimilars within their formulary design. The majority agreed
that using biosimilars in chronic diseases where patients switch from reference brands to
biosimilars would yield the most significant savings. However, the creation of policies
promoting the use of biosimilars first over reference brands would have the most
immediate effect in facilitating the adoption of biosimilars. To support and incentivize
these changes, payers plan to use tier differentials, step edits, and prior authorization
criteria to manage access. Depending on the payer’s plan and patient population, the
type of tools and extent of utilization may vary. For example, a closed formulary system
may conceivably only have access to the biosimilar, whereas an open formulary tiered
plan may have a biosimilar on a lower tier. Payers also commented on the use of
46

biosimilars under the medical benefit. To incentivize usage, payers stated that they
would reimburse physicians at a higher rate for the biosimilar compared to the reference
brand. Although this tactic is presently used to incentivize some brands over others, all
payers noted that management through the medical reimbursement channel is more
challenging than in the pharmacy where most plans have their own pharmacy benefit
manager (PBM). Approximately 1/4 of the plans reported that formulary designs
allowing for preferred biologic brands may not currently exist. As it stands, these plans
currently only have one specialty tier under the pharmacy benefit where both biosimilars
and branded biologics would be placed. Accordingly, payers are conducting ongoing
discussions on methods for facilitating changes that allow for flexibility in incorporating
biosimilars into the formulary when these products become available.
III. Quantitative Results:
Each drug comparison between the reference biologic and the biosimilar consisted of
two discrete choice experiments. The first experiment response required the selection of
product choice (reference or biosimilar); these responses were estimated using a CL
model. The second follow-up experiment response entailed the selection of the
formulary management decision; these responses were estimated using a MNL model.
a. Product Choice Selection
A CL model with fixed effects was used to estimate product choice (‘Biosimilar’ or
‘Reference’). The model was estimated using the following ‘left-out’ levels of each
attribute group: HGH (Therapeutic Area), Launch (Time of Decision), Biosimilar (FDA
Interchangeable), <10% (Cost Differential). These left-out levels were selected as they
represent either baseline or default values, which make comparative interpretations
more intuitive. Nine groups (n=56 observations) were dropped from the first-stage
analysis using the conditional logit model because there was no variation in response
(e.g. the sampled individual selected ‘biosimilar’ for every product comparison
47

administered). As a result, a total of n=205 observations were analyzed. Table 2 shows
the odds ratios (OR) and marginal effects along with respective p- values and
confidence intervals for all tested levels.

Table 2: Odds Ratios and Marginal Effects for Product Selection Using
Conditional Logit Model Estimation

N=205 observations
*Indicates significance at 10% level
**Indicates significance at 5% level
Nine groups and 95 observations were dropped from analysis due to no variation in response at the
individual level
CI, confidence interval; MAB, monoclonal antibody; SD, standard deviation; TNF, tumor necrosis factor
dy/dx indicates change of dummy variable from 0 to 1

The OR (standard deviation) for cost is 1.48 (0.12), indicating that for every %
increase in the cost of the innovator over the biosimilar, the odds of payers preferring the
biosimilar over the innovator is increased by 48%. This effect of price differential is
highly significant. The OR for ‘Time in 12 months’ is 0.09 (0.13) and is significant at the
10% level. Relative to the left- out group ‘Launch (Time of Decision),’ the odds of
Variable
Odds
Ratio
(SD) 95% CI p-value
dy/dx
(SD) 95% CI p-value
TNF Alpha 0.70
(0.53)
0.16-
3.13
0.64
-0.0003
(0.0009)
-0.002-
0.001
0.73
Oncology MAB 2.22
(2.92)
0.17-
29.15
0.54
0.0006
(0.0012)
-0.002-
0.003
0.60
FDA
Interchangeable*
9.79
(13.27)
0.69-
139.44
0.09*
0.0013
(0.0019)
-0.002-
0.005
0.49
Time in 12
months*
0.09
(0.13)
0.01-
1.33
0.08*
-0.003
(0.0052)
-0.01-
0.007
0.56
Time over 12
months
0.70
(0.72)
0.09-
5.35
0.73
-0.00032
(0.0011)
-0.003-
0.002
0.76
Cost** 1.48
(0.12)
1.26-
1.74
0.00**
0.00032
(0.0005)
-0.0006-
0.001
0.49
48

preferring a biosimilar over the reference product is reduced by 91% 12 months after
launch compared to at the launch of the biosimilar. This result implies a preference for
the biosimilar over the innovator at the initial market entry of the product. The OR for
‘FDA Interchangeable’ is 9.79 (SD 13.27) and is also significant at the 10% level. The
odds for preferring a biosimilar are almost tenfold higher when the FDA has designated
the molecule to be ‘interchangeable’ rather than ‘biosimilar.’
The ORs for TNF alpha and oncology mAb, 0.70 (0.53) and 2.22 (2.92), respectively,
were not statistically significant; however, the direction relative to HGH (reference group)
indicates that there may be differential preference for biosimilars by therapeutic class.
Specifically, payers may prefer HGH biosimilars when compared to TNF alpha
biosimilars and oncology mAb biosimilars when compared to HGH biosimilars. Although
the OR for ‘Time over 12 months’ (OR, 0.70; SD 0.72) was not statistically significant,
the interpretation of the direction is consistent with ‘Time in 12 months.’ Specifically, US
payers prefer biosimilars at launch compared to within 12 months or after 12 months of
launch. The marginal effects for all attribute levels were small and not statistically
significant.
b. Formulary Decision Making
In the second discrete choice experiment, a MNL model was used to estimate
formulary choice between biosimilar and reference product. Of the formulary response
choices ‘biosimilar non-formulary,’ ‘biosimilar at parity’, ‘biosimilar preferred (new
patients)’ and ‘biosimilar preferred (current patient switches)’, ‘non-formulary’ was the
left-out reference group since it serves as the default in which payers make no change.
As with the product selection, the same left-out levels in each attribute group served as
the reference for each respective formulary comparison. A total of 300 observations
were analyzed. Tables 3-5 present the ORs and marginal effects as well as the
respective p-values and confidence intervals.
49

i. ‘Biosimilar at Parity’ vs. ‘Biosimilar Non-formulary’
Table 3: Odds Ratios and Marginal Effects for Formulary Selection (‘Biosimilar at
Parity’ vs. ‘Biosimilar Non-formulary’) Using Multinomial Logit Model Estimation

N=300 observations
**Indicates significance at 5% level
CI, confidence interval; MAB, monoclonal antibody; SD, standard deviation; TNF, tumor necrosis factor
dy/dx indicates change of dummy variable from 0 to 1

The OR for cost, 1.21 (0.06), signified that for every % increase in the cost of the
innovator over the biosimilar, the odds that payers are likely to add the biosimilar to the
formulary at parity to the innovator is increased by 21%. This effect of price differential
is highly significant. The marginal effect for cost (-0.02 [0.004]) is also highly significant.
Therefore, with every %-increase in the cost of the innovator over the biosimilar, there is
a 2% decrease in the odds of biosimilar parity access compared to the biosimilar not
being added to formulary. This marginal effect is small but in the opposite direction
compared to the OR. The marginal effects measured here may be sensitive to the
subset of individuals in the sample who would choose not to add the biosimilar to
Variable
Odds
Ratio
(SD) 95% CI P Value
dy/dx
(SD) 95% CI P Value
TNF Alpha 1.18
(0.82)
0.31-
4.58
0.81 0.01
(0.24)
-0.46-0.48 0.97
Oncology MAB** 0.45
(0.60)
0.03-
6.08
0.55 0.33
(0.13)
0.07-0.59 0.01
FDA
Interchangeable
2.01
(2.36)
0.20-
19.99
0.55 0.08
(0.39)
-0.68-0.83 0.84
Time in 12
months**
15.64
(20.50)
1.20-
204.14
0.04 0.18
(0.13)
-0.12-0.38 0.32
Time over 12
months
0.97
(0.94)
0.14-
6.48
0.97 -0.013
(0.18)
-0.38-0.35 0.94
Cost** 1.21
(0.06)
1.09-
1.34
0.00 -0.02
(0.004)
-0.03-
-0.01
0.00
50

formulary in the event that the % cost differences were very low (i.e. less than 10%).
The OR for ‘Time in 12 months’ (15.64 [20.50])is significant at the 5% level and exhibits
a change in sign and magnitude relative to the product preference decision ‘biosimilar’ or
‘reference’ (Table 2). Relative to the left-out group ‘Launch (Time of Decision)’, the odds
of adding a biosimilar at parity to the formulary at 12 months after launch rather than not
adding the biosimilar is increased by nearly16-fold. This finding reinforces the rational
decision that payers strongly prefer biosimilars and will include them in the formulary
rather than deter their use at product launch.
Although the ORs for TNF alpha and oncology mAb (1.18 [0.82] and 0.45 [0.60],
respectively) were statistically significant, the directional change in comparison to the
ORs for product selection may indicate that payers support differential formulary
management for biosimilars by therapeutic class. Specifically, payers may be more
likely to add a TNF alpha biosimilar at parity compared to adding a HGH biosimilar at
parity. This stated preference supports the qualitative findings of the exploratory
research that payers perceive TNF alpha to be the largest biologic class for biosimilar
uptake due to its status as a top budgetary priority. As a result, payers would prefer to
add this biosimilar over HGH which is perceived as a commodity protein with multiple
brands. On the other hand, the results suggest that payers may be less likely to add an
oncology mAb at parity compared to HGH at parity. Likewise, this is consistent with the
exploratory research indicating that due to the complexity of oncology mAbs and
sensitive patient populations, the management of oncology biosimilars may be more
guarded compared to other therapeutic classes.
While the OR for ‘FDA Interchangeable’ (2.01 [2.36]) is not significant; the sign
and magnitude are consistent with the product selection choice indicating that payers
may be more likely to add biosimilars that are interchangeable at parity compared to
those that are not. The direction of the OR for ‘Time over 12 months’ (0.97 [0.94])
51

compared to ‘Time in 12 months’ may indicate that preference may change over time.
Specifically, payers may prefer to manage biosimilars within the first year of availability
compared to launch or after one year of availability. This may reflect the payer’s
perceived ability to maximize cost savings with the adoption of a biosimilar within the
current budget year. The marginal effect for oncology mAb was 0.33 and statistically
significant indicating that an oncology biosimilar relative to a growth hormone biosimilar
is 33% more likely to be covered at parity vs. not being added to the formulary at all.
The marginal effects for TNF alpha, FDA Interchangeable, Time in 12 months, and Time
after 12 months were not statistically significant.
ii. ‘Biosimilar Preferred for New Patients’ vs. ‘Biosimilar Non-formulary’
Table 4: Odds Ratios and Marginal Effects for Formulary Selection (‘Biosimilar
Preferred for New Patients’ vs. ‘Biosimilar Non-Formulary’) Using Multinomial
Logit Model Estimation

N=300 observations
*Indicates significance at 10% level
**Indicates significance at 5% level
Variable
Odds
Ratio
(SD) 95% CI
p
Value
dy/dx
(SD) 95% CI
p
Value
TNF Alpha 1.15
(0.76)
0.31-4.18 0.84 -0.001
(0.51)
-1.00-
0.99
0.99
Oncology MAB 0.12
(0.16)
0.01-1.57 0.11 -0.14
(0.13)
-0.40-
-0.13
0.31
FDA
Interchangeable
1.41
(1.62)
0.15-
13.46
0.77 -0.08
(0.56)
-1.17-
1.02
0.90
Time in 12 months* 9.24
(11.85)
0.75-
114.0
0.08 -0.08
(0.16)
-0.38-
0.23
0.62
Time over 12
months
1.03
(0.92)
0.18-5.94 0.97 0.02
(0.32)
-0.60-
0.64
0.96
Cost** 1.35
(0.07)
1.21-1.49 0.00 0.03
(0.04)
-0.05-
0.10
0.50
52

CI, confidence interval; MAB, monoclonal antibody; SD, standard deviation; TNF, tumor necrosis factor
dy/dx indicates change of dummy variable from 0 to 1

The OR for cost is 1.35 (0.07), indicating that for every % increase in cost of the
innovator over the biosimilar, the odds that payers will prefer the biosimilar for new
patients over not adding the biosimilar to the formulary is increased by 35%. The effect
of price differential on more aggressive formulary management is highly significant. The
OR for ‘Time in 12 months’ (9.24 [11.85]), which is significant at the 10% level, is
consistent in direction and magnitude to the OR generated in the comparison of
‘Biosimilar at Parity’ vs. ‘Biosimilar Non-formulary’ (Table 3). The interpretation is similar
in that relative to the left-out group ‘Launch (Time of Decision)’; the odds of preferring a
biosimilar for new patients at 12 months after launch rather than not adding the
biosimilar at launch is increased by over tenfold.
The sign and magnitude of the ORs for TNF alpha and oncology mAb (1.15
[0.76]) and 0.12 [0.60], respectively) are similar to that of the ORs generated previously
for formulary decision (Table 3). Although not statistically significant, these ORs suggest
that payers may support differential formulary management for biosimilars by therapeutic
class, which is consistent with more aggressive formulary decisions. Specifically, payers
may be more likely to prefer a TNF alpha biosimilar for new patients compared to
preferring a HGH biosimilar for new patients. The result for oncology mAb also suggests
that payers may be less likely to prefer an oncology mAb for new patients compared to
preferring a HGH biosimilar for new patients.
Although not statistically significant, the OR for ‘FDA interchangeable’ (1.41
[1.62]) exhibits a sign and magnitude consistent with the OR produced for formulary
decision (Table 3). None of the estimated marginal effects nor the OR for ‘Time over 12
months’ (1.03 [0.92]) were statistically significant.

53

iii. ‘Biosimilar Preferred for All Patients’ vs. ‘Biosimilar Non-formulary’
Table 5: Odds Ratios and Marginal Effects for Formulary Selection (‘Biosimilar
Preferred for All Patients’ vs. ‘Biosimilar Non-Formulary’) Using Multinomial Logit
Model Estimation

N=300 observations
*Indicates significance at 10% level
**Indicates significance at 5% level
CI, confidence interval; MAB, monoclonal antibody; SD, standard deviation; TNF, tumor necrosis factor
dy/dx indicates change of dummy variable from 0 to 1

The results concerning cost with respect to the most aggressive formulary
management decisions are consistent with the previous findings showing a significant
effect of cost on management. The OR for cost is 1.40 (0.08), indicating that for every %
increase in the cost of the innovator over the biosimilar, the odds that payers are likely to
prefer the biosimilar with all (new and existing) patients over not adding the biosimilar to
the formulary is increased by 40%. The OR for TNF alpha (0.23 [0.76]) was statistically
significant at the 5% level and changed direction relative to the previous ORs for TNF
Variable
Odds
Ratio
(SD) 95% CI
P
Value
dy/dx
(SD) 95% CI
P
Value
TNF Alpha** 0.23
(0.17)
0.06-0.98 0.05 -0.005
(0.74)
-1.46-
1.45
0.99
Oncology MAB** 0.00
(0.00)
0 0.97 -0.25
(0.04)
-0.33-
-0.16
0.00
FDA
Interchangeable
5.98
(7.37)
0.53-
67.04
0.15 0.007
(0.92)
-1.80-
1.80
0.99
Time in 12 months* 8.68
(11.92)
0.59-
128.3
0.12 -0.0006
(0.08)
-0.16-
0.16
0.99
Time over 12
months
0.33
(0.34)
0.04-2.50 0.28 -0.003
(0.44)
-0.87-
0.86
0.99
Cost** 1.40
(0.08)
1.26-1.57 0.00 0.0003
(0.04)
-0.08-
0.08
0.99
54

alpha formulary management. Despite preferring more aggressive formulary
management with TNF alpha biosimilars as previously described, payers are not willing
to prefer this biosimilar for all patients compared to the HGH category. Additionally,
although not statistically significant, the oncology mAb OR (0.00 [0.00]) may be
interpreted to signify a most strict aversion in biosimilar management. The marginal
effect for oncology mAb formulary management is (-0.25 [0.04]) is highly significant.
This supports the results of the OR and can be interpreted such that US payers have a
25% lower odds of preferring oncology mAbs for all patients compared to HGH
biosimilars being preferred for all patients. Together, these findings indicate that with the
adoption of biosimilars, payers may be more likely to require patient switching policies
for HGH than for both TNF alpha and oncology MABs.
The ORs for ‘Time in 12 months’ and ‘Time over 12 months’ (8.68 [11.92] and
0.33 [0.34], respectively) are not statistically significant. Similar to the previous results
(Tables 3 and 4), this may signify that payers will manage biosimilars within the first year
to maximize budget savings; however, this signal may be no longer evident when
considering patient switching policies.
The OR for ‘FDA interchangeable’ (5.98 [7.37]), albeit not significant, may imply
that with aggressive management decisions, payers may be more likely to prefer a
biosimilar for all patients if it is interchangeable. The remaining marginal effects were
not statistically significant.
IV. Attribute Ranking Exercise
At the conclusion of the discrete choice experiments, payers were asked to force
rank the biosimilar attributes included in the product comparisons (e.g. therapeutic area,
time of decision, FDA status, and cost differential). Payers were instructed to order the
four attributes by their importance in formulary management using a scale of 1 to 4,
where 1 indicates the most important criteria for formulary decision making and 4
55

signifies the least important. Due to time constraints, two individuals were unable to
complete the final question; 26 subjects completed the attribute force ranking questions.
Table 6 descriptively summarizes these data by attribute and the respective percentage
of subjects assigning each ranking (1-4).

Table 6: Results from descriptive attribute force ranking exercise

N=300 observations
Question: Please rank how important each attribute is in your formulary decision making for
biosimilars where1 is most important and 4 is least important.

Forty-six percent of the payers selected ‘cost’ as the most important attribute for
biosimilar management, followed by ‘therapeutic area’ (23%), ‘FDA classification’ (19%),
and ‘time’ (12%). The observed rankings show that these payers explicitly acknowledge
the extent to which budget sensitivity influences their decision-making. ‘Therapeutic
area’ was also ranked a high priority, indicating that different therapeutic classes and the
respective patient populations and indications impact how payers perceive biosimilar
opportunities.
Fifty-eight percent of the payers selected ‘time’ as the most ‘least important’
attribute in formulary decision making. The payers indicated that independent of when
the biosimilar is available (e.g. launch, 6-12 months post launch, greater than 12 months
post launch), their preference and opportunity to manage a biosimilar is the roughly the
same. When asked a follow-up question of when ‘time’ may in fact become more
important in the decision-making, many payers responded that a safety signal or
Most Important Least Important
56

possible FDA guidance on safety monitoring would prompt guarded uptake and further
considerations of biosimilar adoption. Alternatively, if the FDA did not establish that the
safety or efficacy profile of the biosimilar as being identical to that of the innovator, US
payers may consider delaying biosimilar management (ie, requiring switching to a
biosimilar).

57

Chapter 5: Discussion
This chapter discusses the adoption of biosimilars by US payers in the context of
how the quantitative and qualitative findings may address the primary and exploratory
hypotheses. Further, limitations with the current research design and model estimation
are examined. Lastly, suggestions for future work related to the overall market impact of
biosimilars including the evaluation of other key stakeholder preferences are considered.
The current research objective is to understand payer access and formulary
decision-making for biosimilars upon market entry. Using two discrete choice exercises
with a sample of US commercial payers, biosimilar product preferences and formulary
management decisions were assessed across a set of hypothetical product choices.
Biosimilar and innovator products were described across four attributes: (1) therapeutic
class, (2) FDA status regarding interchangeability, (3) time of decision, and (4)
percentage cost differential between biosimilar and innovator product. The US payer
interviewee also selected one of four formulary responses for each product choice
ranging in context from no management to most aggressive formulary management: (1)
biosimilar non-formulary, (2) biosimilar added at parity, (3) biosimilar preferred for new
patients, or (4) biosimilar preferred for all (new and existing) patients. Using qualitative
insight from explorative interview research and from previous literature on formulary
management controls in specialty disease areas, the following primary hypotheses were
identified:
I. Primary Hypotheses:
(a) Effect of Price on Product Preference: With larger price differentials offered between
biosimilar and branded biologics, payers are hypothesized to prefer biosimilar products
over the respective innovator products. Results from the quantitative product preference
discrete choice experiment support the effect of price sensitivity. In the CL model, the
58

OR for cost was highly significant and greater than 1 indicating the magnitude and
direction of the price effect are consistent with hypothesized preferences. Based on this
result, US managed care payers in the sample are incentivized to strongly prefer
biosimilars over innovator products with increasing price differences between the two.
Similarly in the qualitative findings, US payers indicated that formulary decision-making
for biosimilars would closely resemble the formulary review for branded products where
cost is a key consideration. Furthermore, this sensitivity to drug cost on drug preference
has been demonstrated with small molecule generics, whereby their uptake has
increased over time and provided cost savings to payers (9). US payers have employed
utilization management controls to incentivize this uptake with both physicians and
patients (4-6). Collectively, prior literature and these findings suggest that biosimilars are
perceived as a lucrative opportunity to generate additional cost savings in growing
specialty drug categories for managed care payers. However, given that the likely cost
differential between biosimilars and innovators is expected to be smaller compared to
small molecule brands and their generics (11, 13) and that standard policies for
automatic pharmacy substitution are lacking, the impact of price on formulary
management may be characterized differently than that of small molecule generics.
(b) Effect of Price on Access Management: Following the hypothesis on price effects on
preference, the second hypothesis focused on how price differentials may influence the
selection of formulary policies payers may use to manage biosimilars compared to
innovator products. With larger price differentials offered between biosimilar and
branded biologics, this hypothesis stated that payers will prefer more aggressive
utilization management that incentivizes biosimilars compared to innovators. Relative to
the policy responses indicated in the second discrete choice experiment (eg, non-
formulary, parity access, biosimilar preferred - new patients, biosimilar preferred – all
patients), this hypothesis asserts that payers are more likely to use formulary placement
59

with increasing biosimilar price discounts. Similar to the established link between price
and product preference in the case of small molecules, literature demonstrates that US
managed care payers have used aggressive management policies (ie, cost-sharing,
prior authorization, or NDC blocks) to drive patient use for generic drugs (9, 12). The
MNL results from the second discrete choice experiment support this hypothesis that
payers may use escalating policies for management with increased price discounts.
Specifically, the cost OR for each formulary management policy comparison (tables 3, 4,
5) are highly significant and greater than 1. Within each comparison, the effect of
increasing the price discount can be interpreted as being associated with increased
payer management. Notably, the values for these ORs (1.21, 1.35, and 1.40) increased
with each formulary comparison. This may support an increasing marginal sensitivity
with respect to price discounting versus a constant rate of marginal sensitivity. In other
words, payers may not be responding proportionally with each additional discount
percentage; there may be an inflection point where aggressive utilization strategies are
endorsed.
(c) Effect of Therapeutic Class on Access Management: Insight from the exploratory
primary research suggested that opportunities for biosimilars may be weighted differently
depending on the therapeutic class to which they belong and the respective patient
populations treated. This evidence is supported with the fact that there are specific
specialty categories that have a significantly higher budget impact compared to others
(9). Furthermore, the management of biosimilars is contingent on current utilization
management in the class of innovator products presently. As a result, a hypothesis
regarding management restrictions varying across therapeutic categories was tested.
Specifically, payer preferences for oncology biosimilars management may be less
aggressive than management for the other two therapeutic classes tested (TNF alpha
and HGH). This finding may be expected considering the sensitivity US payers exhibit
60

with reimbursing oncology treatments. The ORs for the oncology mAb therapeutic
category from the MNL model were less than 1.0, but not statistically significant. This
may indicate that the preference for managing manage an oncology biosimilar over the
innovator was less likely than for managing a biosimilar in the HGH category.
Specifically in the results of the MNL model comparing ‘non-formulary’ to ‘prefer
biosimilar in new patients,’ the OR was 0.12 (p=0.11). Although not statistically
significant, the directional results indicate that US payers are more likely to require new
patients prescribed to HGH to use a biosimilar compared to new patients prescribed to
an oncology biologic. This finding is supported by evidence from recent literature reports
documenting the reticence of US managed care payers in managing oncology therapies
(76). However, recent industry reports have indicated that payers are increasing their
utilization management policies in oncology. This is likely due to increasing budgetary
constraints and new product approvals (77). With the availability of oncology biosimilars
in the future, payers may increase their willingness to manage this class compared to
what is measured in this research. Further research showing how payer management
preferences are evolving within all oncology products will be meaningful for
understanding the degree to which oncology biosimilars may be adopted.
II. Secondary Hypothesis
(d) Explicit vs. implicit attribute weighting: Through the estimation of the ORs in the CL
and ML models, payer preferences for biosimilars are measured implicitly (i.e. observing
a choice selection for a given set of attributes). In many primary research exercises,
explicit measurement of attributes involves exercises where participants are asked to
explicitly rank a set of attributes. The secondary hypothesis examined the consistency
of the explicit and implicit results. Specifically, as indicated by the ORs for ‘cost’
generated in the choice model, payers were hypothesized to be more sensitive to price
61

than the explicit results for ‘cost’ in the ranking exercise. Since ‘cost’ was both highly
significant in the implicit models and identified as the highest ranking factor for the
majority of payers, the hypothesis of a possible difference was not substantiated with
respect to cost. In contrast, there was a discrepancy with respect to the ‘time’ attribute
between the implicit results and explicit rankings. In the discrete choice estimation, the
OR for ‘Time in 12 months’ was significant indicating that payers were more likely to
prefer a biosimilar over an innovator biologic at 12 months after launch compared to
immediately at product launch. This finding is consistent with the time period required
for the formal annual P& T review which is conducted for all newly–approved products.
During the explicit ranking exercise; however, the majority of payers indicated that ‘time’
is the least important attribute in formulary decision making. This discrepancy could be
due to the difference between the definitions of time as criteria vs time as a step in the
decision-making process. As part of the discrete choice exercise, ‘time’ is used to
indicate a period correlating with the formulary review cycle (i.e. at FDA approval and
launch, 6-12 months later are the first formal review, and after 12 months for subsequent
review). Payers may not directly consider how their decision-making is influenced by
these different time points; however, payers may be more likely to prefer biosimilars by
the first year of availability because this is the period in the process where they are more
likely to make a change to formulary management that impacts all products within the
class. Therefore, ‘time’ is part of the planning and reviewing process in which they make
their decisions. In contrast, in the explicit ranking exercise, payers may be considering
how all the attributes refer to the product itself and not to their decision-making process.
For example, payers may interpret ‘cost’ as the price differential offered by the biosimilar
and ‘time’ as how long the biosimilar has been available in the market. With the latter
interpretation, ‘time’ may not be considered highly important since payers are eager to
adopt biosimilars to capture immediate budget savings. The convention and process of
62

their formulary review, however, are the variables being measured and described in the
discrete choice experiment.
III. Additional payer policy considerations
Both the qualitative and quantitative analyses indicated that US managed care
payers are anticipating the opportunities for budget savings offered by biosimilars.
Specifically, the operational ways in which payers manage biosimilar products and drive
uptake may vary as a function of price and time. Although not significant, there were
trends in the differences among the three therapeutic classes tested. Despite
opportunities for capturing the largest cost savings by converting patients to a less
expensive biosimilar, payers may not be willing to require patient switching at the risk of
clinical uncertainty. These data support the notion that US payers may be more likely to
treat biosimilar products as less expensive branded molecules rather than like small
molecule generics.
In this research, payers demonstrated cost sensitivity with respect to product
preference as well as the degree of aggressive formulary management policies.
Although the current study only examined the choice between the innovator and one
biosimilar product as representing the final negotiation choice, the likely scenario payers
will encounter will be a dynamic situation involving negotiations with both manufacturers.
As a result, payers will leverage discount prices for one product with the competitor
manufacturer in an attempt to reap the highest cost savings. With this strategy, it is
conceivable that payers may negotiate savings with the innovator product and decide
not to prefer the biosimilar. Ultimately, the benefit of biosimilars to payers is the
opportunity for competition and cost savings, whether this is provided through switching
63

patients to biosimilars or pressuring innovator manufacturers into offering significant
discounts for the branded biologic.
IV. Research Limitations
Several limitations are noted in the current research experiment and analysis.
These limitations relate to either technical or methodological considerations or
applicability to real-world payer decision-making.
a. Methodological Limitations
The research design included a limited number of US managed care payer
subjects and consisted of a convenient sample. These individuals were selected based
on a set of pre-defined criteria but may not be representative of all payers and plans.
Specifically, payers who manage state Medicaid patient populations are under-
represented in the current sample; as such, the observed results may not be indicative
of biosimilar adoption in this population. Due to the more pronounced budget constraints
in managing this population, biosimilar preferences and management with respect to
cost may bear a larger effect than what has been presented here for this segment of
payers.
Another study design limitation arises from the attribute selection and level
assignment, which were largely influenced by the pilot research responses and the
literature outlining formulary decision-making criteria for other therapeutic categories
(6,7,8,9). Not all attributes, including biosimilars of different therapeutic class, were
tested. Although this research included three therapeutic classes that appeared to
provide the largest payer benefit either due to potential budget savings (i.e. TNF alpha
and oncology mAbs) or limited clinical risk (HGH), these classes are not exhaustive.
Payers may exhibit different preferences and management behaviors for other
64

therapeutic molecules. Additionally, the product descriptions in the survey may be
considered vague relative to the product profiles that would be presented to P&T
committee members in the real world. However, in the absence of an FDA-approved
biosimilar at the time of the research fielding, profiles with attribute levels that
approached those expected by payers for evaluation were provided.
Additionally, the subset of comparisons was selected to test main effects only.
With this specification, no interactions between the attributes have been tested. Given
that the research questions are focused on overall biosimilar uptake spanning across
several potential scenarios in different therapeutic classes and patient populations, it is
reasonable that main effects may be sufficient to describe the attitudinal behaviors.
Interaction terms may be suitable for a follow-on research study that entails specific
brand comparisons where the scope of the research is more narrowly defined. The
selection of the conditional logit and multinomial logit models poses another potential
methodological limitation. These specifications are not as flexible as other model
choices, such as the random parameters logit model which allows for heterogeneity in
the measurement of individual taste weights. Considering the computational and
programming complexities required for this form and estimation technique, the
conditional logit and multinomial logit models were selected and considered sufficient to
address the research hypotheses. Further discussion of the trade-offs associated with
model selection is provided in ‘Chapter 5: Methods.’
b. Applicability Limitations
An important limitation in using SP methods to elicit choice behaviors is the
caveat that decisions made by respondents are not always consistent with revealed
market behavior. The response behaviors elicited through the discrete choice study
65

here revealed preference under hypothetical circumstances, which may or not be
characteristic of the true future market conditions for biosimilars in the US. With
biosimilar market entry being approximately near term in the US, revealed preference
modeling is one of the few methods available to objectively characterize these decision-
making preferences and the management responses payers may be likely to employ.
Despite the limitations in accounting for true market constraints, these data may be
appropriate for predicting changes in organizational decision-making since these trends
are slow to change over time. US payers may not be faced with exactly the same
market conditions as described in the survey; however, from a practical stance, SP
modeling provides an objective approach to approximating influential factors within the
market before biosimilars become available.
Since the time of the research fielding, greater experience with biosimilars has
unfolded in several European countries which may influence the expectations of
managed care payers in the US. Specifically, the adoption of biosimilars in Europe may
have contributed to evolving perceptions of the availability of biosimilars in the US. The
most significant experience in Europe that may augment US payer expectations
concerns the price discounts observed to date. Price discounts in certain markets have
exceeded the expectation of 10-30% suggested by several studies (11,13). In July
2015, a large hospital group in France awarded a tender to a TNF biosimilar infliximab
(brand: Remicade) with a 45% discount to the brand (78). In a Norway hospital tender in
January of 2015, another manufacturer offered the drug procurement agency a 69%
discount for their Remicade biosimilar (79). In addition to providing real-world
experience of the budget savings opportunities for payers, these significant discounts
have prompted reactions from other biosimilar manufacturers as well as investors as to
whether the wide-spread availability and manufacturing of biosimilars is sustainable at
66

these prices. Specifically, as the profit margins decrease with increasing discounts in
Europe, the dynamics of how manufacturers will compete in the US are subject to
change.
In addition to the pricing observed within the tender markets, several European
health technology assessment (HTA) organizations have offered guidance on reviewing
biosimilars for reimbursement. In January 2015, NICE issued an updated statement
indicating that there is no requirement for a dossier submission at biosimilar product
launch. Furthermore, biosimilar manufacturers will be invited to participate during a
review of an innovator drug (80). Similarly, the Scottish Medicines Consortium (SMC)
issued guidance in May of 2015 asserting that it would no longer routinely assess
biosimilar medicines on the basis of a full submission. If the reference product has been
accepted by SMC for the same indication prior to January 31, 2002, no full submission is
required and the biosimilar would be evaluated (81). The implication of these guidelines
is that contrary to the hurdles required for branded biologics, biosimilar products may be
reviewed and endorsed for reimbursement without the requirement of a data submission
package. Similarly in the US, managed care payers may employ guidance and policies
during formulary review whereby biosimilars are added to formulary without the need for
a full submission package (i.e. AMCP value dossier). Further reduction in the barriers to
entry could potentially change the timing of biosimilar formulary decisions from those
studied in this research.
V. Future Research
Given the results and limitations with the current research study on biosimilar
preference in US managed care payers, key opportunities for related research studies
may expand the insights into payer adoption. With recent experience and uptake of
67

biosimilar TNFs in the hospital channel, conducting another preference study involving a
broader set of US payers not limited to managed care would be of interest. A sample
comprised of hospital payers as well as state Medicaid payers would provide a more
comprehensive picture of payer preferences regarding patient access to biologics in the
US and may potentially demonstrate greater price sensitivities than those measured in
the current study.
Considering the experience in biosimilar pricing and uptake observed in Europe,
a discrete choice experiment using a sample of European payers to compare how risk
attitudes and pricing sensitivities can vary may also be of interest.
This research generated an additional hypothesis with respect to how payer
decision-making for biosimilars compares to the established decision-making for small
molecule generics. Another study might include discrete choice comparisons for specific
biosimilar and innovator molecules as well as small molecule brands and their generics.
Comparing the ORs for selection and management may assist in determining how risk
tolerances differ. Additionally, since such an exercise would include comparisons
specific to molecules and therapeutic categories as in the present study, testing and
measuring interaction terms between the attributes may be also feasible.
To gain a broader understanding of how policy implications can influence patient
access to biosimilar therapies in the US, it is critical to capture the conditional
prescribing behaviors of physicians. Many payers are trained clinicians and also call on
physicians to provide recommendations in specific disease areas pending new policy
changes or coverage decisions for new drugs. Physicians who are risk-averse and
willing to challenge payer policies on biosimilars may create hurdles and resistance
towards payer utilization management efforts. A discrete choice experiment including
68

physicians across specialties may reveal differences in risk attitudes toward biosimilars
and provide insight into how these attitudes may influence payer policies and patient
access.
VI. Conclusion
With the trend of increasing drug expenditures in specialty biologics in the US,
managed care payers are anticipating the introduction of biosimilars that offer lower-cost
options with clinical certainty compared to the reference brand. This research examined
how a convenience sample of US managed care payers might approach these novel
therapies with regards to product preferences and ultimately, policy changes for
formulary placement. Engaging in a two-stage discrete choice exercise, payers were
asked to respond to a product comparison of an innovator and biosimilar therapy based
on the following attributes: therapeutic class, time of decision, FDA status, and cost
differential between the brand and biosimilar. The findings showed that payers are
highly cost-sensitive such that greater differences in price between the biosimilar and
innovator product will increase the odds of preference for the biosimilar. ‘FDA status’
and ‘time to decision’ also showed trends toward significance. The odds for payer
preference were also increased for biosimilars approved with a more rigorous clinical
designation of ‘interchangeable’ status. The results also suggest that due to the
organizational process and timelines for drug review, payers are more likely to prefer a
biosimilar within the first 12 months of launch compared to at the time of product launch.
In the follow-up policy question, payers responded to a question using the same
comparison with respect to endorsing a specific management decision. The payers’
preferences for more aggressive management, including requiring the biosimilar for all
new and existing patients, were largely influenced by the price discount. Although not
consistently significant in all comparisons, signals of risk attitude differences across
69

therapeutic classes of biologics were detected. Payers preferred adopting a biosimilar
with an increasing price concession, but were less likely to switch patients to a TNF
alpha and oncology biosimilar than to a biosimilar HGH. Collectively, these research
findings showed that payers are orienting themselves to novel biosimilar agents by
balancing the perceived budget savings with tolerance for some degree of clinical
uncertainty.

70

References

(1) Organization for Economic Co-operation and Development 2010, "OECD Health
Data", OECD Health Statistics Database accessed Oct 2015
(2) IMS Institute for Healthcare Informatics. Medicine use and shifting
costs of healthcare: A review of the use of medicines in the United States in 2013.
April 2014
(3) Humphreys A. 2004. Difficult path for biogenerics. MedAdNews, December 2004
(4) Goldman DP, Joyce GF, Lawless G, Crown WH, Willey V. Benefit design and
specialty drug use. Health Aff. (Millwood). 2006;25(5):1319-31
(5) Willey VJ, Kopenski F, Lawless GD. Beyond the myths: finding benefit design
solutions that address the true costs of high healthcare use. Am J Manag Care.
2008;14(8):S252-S263
(6) AARP Strategic analysis and intelligence (SAI) report: The tier 4 phenomenon:
shifting the high cost of drugs to consumers. March 9, 2009. Available at:
http://assets.aarp.org/rgcenter/health/tierfour.pdf. Accessed October 1, 2009
(7) Polinski JM, Mohr PE, Johnson L:Impact of Medicare Part D on Access to and
Cost Sharing for Specialty Biologic Medications for Beneficiaries With Rheumatoid
Arthritis. Arthritis & Rheumatism Vol 61, No. 6, June 15, 2009 pp 745-754
(8) Green CJ, Maclure M, Fortin PM, Ramsay CR, Aaserud M, Bardal S.
Pharmaceutical policies: effects of restrictions on reimbursement (Review). The
Cochrane Collaboration, Issue 8. June 2010
(9) IMS Health, Plantrak Copay, Formulary Focus, Dec 2010. Reference: The Use
of medicines in the United States: Review 2010 April 2011
(10) Patient Protection and Affordable Care Act of 2010, Pub. L. no. 111-148, 124
Stat 1025 (2010). Print.
71

(11) Grabowski H, “Follow-on biologics: data exclusivity and the balance between
innovation and competition” Nat Rev Drug Discov. 2008 Jun;7(6):479-88.
(12) Saha A, Grabowski H, Birnbaum H, Greenberg P. Bizan O. “Generic
Competition in the US Pharmaceutical Industry.” Int. J. of the Economics of Business
Vol. 13, No. 1, February 2006, pp.15-38
(13) Congressional Budget office Cost Estimate: Biologics Price Competition and
Innovation Act. June 2008
(14) Nichol MB, Knight TK, Epstein MA, Honda D, Tretiak R. Opinions Regarding the
Academy of Managed Care Pharmacy Dossier Submission Guidelines: Results of a
Small Survey of Managed Care Organizations and Pharmaceutical manufacturers.
Journal of Managed Pharmacy May 2007 Vol. 13, No. 4
(15) Lancaster, K. A new approach to consumer theory. Journal of Political Economy
74: 132-57, 1966
(16) Rosen, S. Hedonic prices and implicit markets: product differentiation in pure
competition. Journal of Political Economy 82(1):34-55, 1974
(17) Louviere J, Hensher D, Swait J. Stated Choice Methods: Analysis and
Applications. Cambridge, UK: Cambridge University Press, 2000
(18) Bridges J, Kinter E, Kidane L, et al. Things are looking up since we started
listening to patients: recent trends in the application of conjoint analysis in health 1970-
2007. Patient 2008;1:273-82
(19) Ryan M, Gerard K. using discrete choice experiments to value health care
programmes: current practice and future research reflections. Appl health Econ Health
Policy 2003;3:79-89
(20) Lancsar E, Louviere J. Conducting discrete choice experiments to inform
healthcare decision making: a user’s guide. Pharmacoeconomics 2008;26:661-77
72

(21) Ryan M, Farrar S. Using conjoint analysis to elicit preferences for health care.
BMJ 2000; 320:1530-3
(22) Bridges JFP et al. Conjoint Analysis Applications in Health-A Checklist: A Report
of the ISPOR Good Research Practices for Conjoint Analysis Task Force. Value in
Health 14 (2011) 403-413
(23) Weston A, Fitzgerald P. Discrete choice experiment to derive willingness to pay
for methyl aminolevulinate photodynamic therapy versus simple excision surgery in
basal cell carcinoma. Pharmacoeconomics 2004;22:1195–208
(24) Mühlbacher AC, Lincke H-J, Nübling M. Evaluating patients’ preferences for
multiple myeloma therapy, a discrete-choice experiment. Psychosoc Med
2008;5:Doc10
(25) Bridges J, Selck F, Gray G, et al. Common avoidance and determinants of
demand for male circumcision in Johannesburg, South Africa. Health Policy Plan 2010;
Oct 20
(26) Phillips KA, Maddala T, Johnson FR. Measuring preferences for health care
interventions using conjoint analysis: an application to HIV testing. Health Serv Res
2002;37:1681–705
(27) Beusterien KM, Dziekan K, Flood E, et al. Understanding patient preferences for
HIV medications using adaptive conjoint analysis: feasibility assessment. Value Health
2005;8:453–61
(28) Coast J, Salisbury C, de Berker D, et al. Preferences for aspects of a
dermatology consultation. Br J Dermatol 2006;155:387–92
(29) King MT, Hall J, Lancsar E, et al. Patient preferences for managing asthma:
results from a discrete choice experiment. Health Econ 2006; 16:703–17
73

(30) Peacock S, Apicella C, Andrews L, et al. A discrete choice experiment of
preferences for genetic counseling among Jewish women seeking cancer genetics
services. Br J Cancer 2006;95:1448–53
(31) Regier DA, Ryan M, Phimister E, et al. Bayesian and classical estimation of
mixed logit: an application to genetic testing. J Health Econ 2009;28;598–610
(32) Roux L, Ubach C, Donaldson C, et al. Valuing the benefits of weight loss
programs: an application of the discrete choice experiment. Obesity Res Clin Pract
2004;12:1342–51
(33) Hauber AB, Johnson FR, Sauriol L, et al. Risking health to avoid injections:
preferences of Canadians with type 2 diabetes. Diabetes Care 2005;28:2243–5
(34) Johnson FR, Manjunath R, Mansfield CA, et al. High-risk individuals’ willingness
to pay for diabetes risk-reduction programs. Diabetes Care 2006;29:1351–56
(35) Marshall DA, Johnson FR, Kulin NA, et al. How do physician assessments of
patient preferences for colorectal cancer screening tests differ from actual
preferences? Comparison in Canada and the United States using a stated-choice
survey. Health Econ 2009;18:1420–39
(36) Marshall DA, McGregor E, Currie G. Measuring preferences for colorectal cancer
(CRC) screening—what are the implications for moving forward? Patient 2010;3:79–89
(37) Wittink MN, Cary M, TenHave T, et al. Towards patient-centered care for
depression: conjoint methods to tailor treatment based on preferences. Patient
2010;3:145–57
(38) Hauber AB, Johnson FR, Mohamed AF, et al. Older Americans’ risk-benefit
preferences for modifying the course of Alzheimer’s disease. Alzheimer Dis Assoc
Disord 2009;23:23–32
(39) Shumway M. Preference weights for cost-outcome analyses of schizophrenia
treatments: comparison of four stakeholder groups. Schizophr Bull 2003;29:257–66
74

(40) Lee JY et al. Impact of Pharmacometric Analysis on New Drug Approval and
Labelling Decisions: A Review of 198 Submissions Between 2000 and 2008. Clin
Pharmacoklnet 2011; 50(10):627-635
(41) Tominaga T, Asahina Y, Uyama Y, Kondo T. Regulatory Science as a Bridge
Between Science and Society. Clinical Pharmacology & Therapeutics Vol 90 No 1
July 2011
(42) Simon GE, Psaty BM, Hrachovec JB, Mora M. Principles For Evidence-Based
Drug Formulary Policy. Journal of General Internal Medicine Vol 20, Issue 10 964-
968, Oct 2005
(43) Carroll N. How Effectively Do managed Care Organizations Influence
Prescribing and Dispensing Decisions? American Journal of Managed Care
2002;8:1041-1054
(44) Jacobson D. Effective Strategies for Managing Pharmacy Benefits. Healthcare
Financial management. March 2001
(45) Olsen B. Approaches to Pharmacy Benefit Management and the Impact of
Consumer Cost Sharing. Clin Therapeutics 2003; 25:250-272
(46) Huskamp H, Deverka P, Epstein A, Epstein R et al. The Effect of incentive-
based formularies on prescription-drug utilization and spending. The New England
Journal of Medicine. Vol 349 Issue 23 2224, 2003
(47) Sica J. Managing prescription drug costs. Employee Benefits Journal; Mar
2001; 26, 35-40
(48) Huskamp H, Epstein A, Blumenthal D. The Impact of a National Prescription
Drug Formulary On Prices, Market Share, and Spending: Lessons for Medicare?
Health Affairs, 22, no. 3 (2003): 149-158
75

(49) Johnson FR, Backhouse M. Eliciting Stated Preference for Health-Technology
Adoption Criteria Using Paired Comparisons and Recommendation Judgments. Value
in Health. Vol 9 Number 5. 2006; 303-311
(50) Keeney RL. Decision Analysis: An Overview. Operations Research, Vol 30, No.
5 (Sept-Oct, 1982) pp. 803-838.
(51) Gilbert RJ, Richels RG. Reducing Uranium Resource Uncertainty: Is it Worth
the Cost? Resources and Energy 1981. 3,13-38.
(52) Howard et al. The Decision to See Hurricanes. Science 176, 1191-1202. 1972
(53) North et al. Planning Wildfire Protection for the Santa Monica Mountains. Fire J.
69, No 1 1975
(54) Edwards W. Reflections on and Criticisms of a Political Multiattribute Utility
Analysis. Mathematical Frontiers of Behavioral Policy Sciences. L. Cobb and RW
Thrall (eds) Westview Press, Boulder, Col. 1980
(55) North D.W., Merkhhofer M.W. A Methodology for Analyzing Emission Control
Strategies. Computers Opn. Res. 3, 185-207. 1976
(56) von Winterfeldt D. Setting Standards for Offshore Oil Discharges: A Regulatory
Decision Analysis. Opns. Res. 30, 867-886, 1982
(57) Keefer DL, and Kirkwood CW. A Multiobjective Decision Analysis: Budget
Planning for Product Engineering. J. Opnl. Res. Soc. 29, 435-442.1978
(58) Smallwood RD, Morris PA. A Task Force Decision Analysis. Opns. Res. 18, 60-
80.1980
(59) Stillwell et al. Evaluating Credit Applications: A Validation of Multiattribute Utility
Techniques against a Real World Criterion, SSRI Research Report 80-1, Social
Science Research Institute, University of Southern California, Los Angeles. 1980
76

(60) Karren R and Barringer M. A review and analysis of the policy-capturing
methodology in organizational research: Guidelines for research and practice.
Organizational Research Methods 5, 337-38. 2002
(61) Cooksey, RW. Judgment Analysis: Theory, methods, and applications. San
Diego , CA: Academic Press. 1996
(62) Rossi PH, Anderson AB. The factorial survey approach: An Introduction. In P.H.
Rossi & S.L. Nock (Eds.), Measuring Social Judgments: A factorial survey approach .
Beverley Hills, CA: Sage. 1982
(63) Sanchez JL, Levine EL. Determining important tasks within jobs: A Policy
Capturing approach. Journal of Applied Psychology. 74, 336-342. 1989
(64) Graves LM & Karren RJ. Interviewer decision processes and effectiveness: An
experimental policy-capturing investigation. Personal Psychology, 45, 313-340. 1992
(65) Rynes S, Lawler J. A policy-capturing investigation of the role of expectancies in
decisions to pursue job alternatives. Journal of Applied Psychology, 68,620-631.1983
(66) Viswesvaran C, Barrick MR Decision-making efforts on compensation wages:
Implications for market wages. Journal of Applied Psychology,77,588-597. 1992
(67) Brush et al. Doctoral Education in the field of entrepreneurship. Journal of
Management, 29,309-331. 2003
(68) Shane S, Venkataramann N. The promise of entrepreneurship as a field of
research. Academy of Management Review. 25,217-226. 2000.
(69) Baron R. Opportunity recognition as pattern recognition: how entrepreneurs
“connect the dots” to identify new business opportunities. Academy of Management
Perspectives, 20,104-119. 2006
(70) Norton W, Moore W. The influence of entrepreneurship risk assessment on
venture launch or growth decisions 2006,
77

(71) Palich L, Bagby D. using cognitive theory to explain entrepreneurial risk-taking:
Challenging conventional wisdom. Journal of Business Venturin, 10, 425-438. 1995.
(72) Janney J, Dess G. The risk concept for entrepreneurs reconsidered: New
Challenges to the conventional wisdom. Journal of Business Venturing, 21, 385-400.
2006.
(73) Venkataraman S. The distinctive domain of entrepreneurship research: An
Editor’s perspective. In J. Latz & R Brockhaus (Eds.), Advance in entrepreneurship,
frim emergence, and growth (Vol 3, pp.119-138). Greenwich, CT: JAI. 1997.
(74) Rousseau D, Sitkin S, Burt R, Camerer C. Not so different afterall: A cross-
discipline review of trust. Academy of Management Review, 23, 393-404. 1998.
(75) Schoder D, Haenlein M. The relative importance of different trust constructs for
sellers in the online world. Electronic Markets, 14, 48-57. 2004
(76) Kenney JT. Payers' Management of Oncology Drugs: Opportunities and
Challenges. Am Health Drug Benefits. 2014 May; 7(3): 123–124.
(77) Magellan Rx Management. Medical Pharmacy Trend Report 2013. Fourth
Edition. 014. http://magellanrxinsights.com/wp-
content/uploads/dlm_uploads/2014/04. Magellan-Rx-Management_Trend-
Report_2013.pdf.
(78) Hirschler B and Callus A. (2015, July 3). Hospira wins French biosimilar
drug tender at 45 percent discount. Reuters. Retrieved from
http://uk.reuters.com/article/2015/07/03/uk-france-biosimilars-exclusive-
idUKKCN0PD1W320150703.
(79) Char P. (2015, March 6). Copy me: Europe shows U.S. a path to
cheaper biotech drugs. Reuters. Retrieved from
78

http://uk.reuters.com/article/2015/03/06/uk-health-biosimilars-insight-
idUKKBN0M20IP20150306.
(80) National Institute For health and Care Excellence (NICE). (2015, Jan 6).
NICE’s biosimilars position statement. Retrieved from
https://www.nice.org.uk/Media/Default/About/what-we-do/NICE-guidance/NICE-
technology-appraisals/biosimilars-statement.pdf
(81) Scottish medicines Consortium (SMC). (2015, May). Biosimilar
Medicines. Retrieved from https://www.nice.org.uk/Media/Default/About/what-
we-do/NICE-guidance/NICE-technology-appraisals/biosimilars-statement.pdf

79

Appendix A: Instrument Key
Attributes
No. of
Levels
Levels Description
Therapeutic
Class
3
TNF alpha inhibitors Inhibits inflammatory
responses associated
with autoimmune
disorders
Human Growth
Hormone
Peptide hormone used to
great children's growth
disorders and adult
hormone deficiency
Oncology mAbs Target mechanisms
against malignant cells for
cancer treatment
Time of Decision
*no evidence of
difference in safety
indicators between
two products*
3
Launch market entry and approval
of
biosimilar/interchangeable
product

6-12 months Preliminary P&T review
(6-12 months after
approval)
>12 months/Annual
review
Annual class review
following initial review
Biosimilar FDA approved drug
demonstrating clinical and
non-clinical evidence to
support similar safety and
efficacy compared to
reference
Interchangeable FDA approved drug
demonstrating higher
threshold of clinical and
non-clinical evidence to
support similar safety and
efficacy compared to
reference with no
difference in any given
patient
Cost Differential
*innovator product
will always have
higher price
4
<10%
Differential spread in price
for innovator compared to
new entrant
10-20%
20-30%
>30%

80

Appendix B: Example of Two-step Discrete Choice Example

Abstract (if available)

Abstract With the trend of increasing drug expenditures in specialty biologics in the US, managed care payers are anticipating the introduction of biosimilars, lower-cost therapeutic options with clinical certainty compared to the reference brands. This research examined how a convenience sample of US managed care payers might manage access to these novel therapies with regards to product preferences. Engaging in a two-stage discrete choice exercise, payers were asked to respond to a product comparison of an innovator and biosimilar therapy based on the following attributes: therapeutic class, time of decision, FDA status, and cost differential between the brand and biosimilar. The results showed that payers are highly cost-sensitive such that greater differences in price between the biosimilar and innovator product will increase the odds of preference for the biosimilar. ‘FDA status’ and ‘time to decision’ also showed trends toward significance. The odds for payer preference were also increased for biosimilars approved with a more rigorous clinical designation of ‘interchangeable’ status. The results also suggest that due to the organizational process and timelines for drug review, payers are more likely to prefer a biosimilar within the first 12 months of launch compared to at the time of product launch. In the follow-up policy question, payers responded to a question using the same comparison with respect to endorsing a specific management decision. The payers’ preferences for more aggressive management, including requiring the biosimilar for all new and existing patients, were largely influenced by the price discount. Although not consistently significant in all comparisons, signals of risk attitude differences across therapeutic classes of biologics were detected. Payers preferred adopting a biosimilar with an increasing price concession, but were less likely to switch patients to a TNF alpha and oncology biosimilar than to a biosimilar HGH. Collectively, these research findings showed that payers are orienting themselves to novel biosimilar agents by balancing the perceived budget savings with tolerance for some degree of clinical uncertainty.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

The impact of treatment decisions and adherence on outcomes in small hereditary disease populations

PDF

Essays on access to prescription drugs

PDF

Evaluating approaches to reduce inappropriate antibiotic use in the United States

PDF

Three essays on comparative economic and clinical outcomes of antipsychotics using retrospective claims data

PDF

Delivering better care for children with special health care needs: analyses of patient-centered medical home and types of insurance

PDF

The causal-effect of childhood obesity on asthma in young and adolescent children

PDF

Effects of a formulary expansion on the use of atypical antipsychotics and health care services by patients with schizophrenia in the California Medicaid Program

PDF

Understanding primary nonadherence to medications and its associated healthcare outcomes: a retrospective analysis of electronic medical records in an integrated healthcare setting

PDF

The value of novel antihyperlipidemic treatments in the U.S. healthcare system: Reducing the burden of cardiovascular diseases and filling the gap of low adherence in statins

PDF

Impact of pharmacy-based transitional care on healthcare utilization and costs

PDF

An evaluation of risk predictors for disease progression in Veteran Affairs patients with chronic hepatitis C

PDF

Essays in opioid use and abuse

PDF

Essays in pharmaceutical and health economics

PDF

Real-time controls in revenue management and service operations

PDF

Use of natural colors: experience and views in pharmaceutical and dietary supplement industries

PDF

Pharmacokinetic and pharmacodynamic optimization of CFTR modulator therapy to mitigate potential drug interactions and adverse events in people with cystic fibrosis

PDF

Acculturation team-based clinical program: pilot program to address acculturative stress and mental health in the Latino community

Asset Metadata

Creator Schwartz, Elizabeth Lydia (author)

Core Title US biosimilar market entry and uptake: US commercial payer preferences in access decision making

School School of Pharmacy

Degree Doctor of Philosophy

Degree Program Pharmaceutical Economics and Policy

Publication Date 11/16/2015

Defense Date 10/07/2015

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag biologics,biosimilar pricing,biosimilars,discrete choice,follow-on biologics,managed care,market access,North America,OAI-PMH Harvest,preferences

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Nichol, Michael B. (committee chair), Doctor, Jason (committee member), Joyce, Geoffrey (committee member)

Creator Email elschwartz426@gmail.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c40-199646

Unique identifier UC11279099

Identifier etd-SchwartzEl-4041.pdf (filename),usctheses-c40-199646 (legacy record id)

Legacy Identifier etd-SchwartzEl-4041.pdf

Dmrecord 199646

Document Type Dissertation

Format application/pdf (imt)

Rights Schwartz, Elizabeth Lydia

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA