REPORT

A Case Study on
Rheumatoid Arthritis

Tom Baker, MPA

Abstract the low incidence of most of the targeted

At a time when many managed care organizations conditions has limited the ability of most
increasingly shift costs to patients through tiered for- health plans to negotiate favorable volume-
mularies and widening copay differentials, biologic based discounts, effectively exposing them
agents represent a significant clinical and financial
to the full brunt of biologic prices.
challenge unlikely to be managed optimally with
Faced with a burgeoning biologics pipe-
tiered formularies and greater patient cost sharing.
The information discussed in this article is intend- line, many managed care organizations
ed for healthcare professionals involved with (MCOs) have begun to apply traditional
rheumatoid arthritis therapy, including but not limit- pharmacy management tools and tech-
ed to physicians in both the inpatient and outpatient niques to biologic agents. Given the antici-
setting, and for other managed care professionals, pated price tags of new biologic agents,
including medical directors, pharmacy directors, many MCOs anticipate using tiered formula-
long-term care decision makers, nurses, pharmacists, ries, cost sharing, percentage copays, and
and case managers. other strategies to limit cost exposure.
(Am J Manag Care. 2003;4:S87-S98) However, because these newer agents offer
substantial efficacy gains, their evaluation
within the limited parameters of small
molecule cost-benefit may undervalue their
Biologic Agents in the Era of potential impact.
Consumer Ascendancy Although many physicians and managed
A review of manufacturer pipelines iden- care professionals recognize rheumatoid
tifies approximately 400 biologic and arthritis (RA) as a complex and compara-
injectable therapies currently in develop- tively costly condition to treat, many believe
ment. Many of these products will offer that optimal pharmacologic management of
patients remarkable gains in quality of the disease may not be achieved using stan-
life (QOL), life expectancy, and produc- dard managed care cost containment strate-
tivity. Most will be expensive, reflecting gies. In particular, current trends to shift
the considerable technological achieve- costs to patients may place expensive bio-
ments required to develop safe, efficacious logic agents beyond the reach of many
biopharmaceuticals. patients.
Until only recently, most biologic agents At the same time, health plan decision
targeted conditions characterized by com- makers reinforced the necessity of a clear
paratively low incidence in the general pop- product value proposition, particularly as
ulation. As a consequence, the significant biologic agents come to represent an even
acquisition costs associated with the prod- larger share of the available treatment arma-
ucts have had a relatively minor per mem- mentarium. Expensive biologic agents may
ber per month impact, particularly when offer unique advantages in efficacy and tol-
compared with any of the more familiar erability, but decision makers cautioned
mass-marketed small molecule categories, that their MCOs lack effective frameworks
such as proton pump inhibitors, selective for evaluating them, particularly at prices
serotonin reuptake inhibitors, and nonse- several times higher than comparable oral
dating antihistamines. On the other hand, agents. Although efficacy advantages will

VOL. 9, NO. 4, SUP. THE AMERICAN JOURNAL OF MANAGED CARE S87

REPORT
likely be substantial, and may translate into copay differentials, biologic agents present a
tangible economic benefits, the managed clinical and financial challenge unlikely to
care systems currently in place may be ill be managed optimally with tiered formula-
suited for assessing and managing the bene- ries and greater patient cost sharing. Many
fits of biologic agents. In addition, plans face believe that although the entry of a third
the added challenge of trying to manage biologic agent (adalimumab)*1 into the mar-
costs for comparatively small populations. ket for disease-modifying antirheumatic drugs
Further, these issues must be considered (DMARDs) will have an impact on their MCOs,
in the context of growing patient and con- they expressed greater concern about their
sumer involvement in healthcare decisions. ability to manage these and other biologic
At a time when health plans increasingly agents effectively, thereby realizing the ben-
offer more open pharmacy benefits in efits associated with greater efficacy and bet-
exchange for greater patient cost sharing, ter targeting.
the use of tiered benefits—and tiered out-of- Currently, manufacturers have about 400
pocket payments—presents significant chal- biotech products in the pipeline, the major-
lenges. For example, even a $50 top-tier ity of which will cost in excess of $12 000
copayment represents a comparatively tiny per year. MCOs, currently moving away
percentage of the total price, thus mitigating from supply-side cost control strategies, fear
its effect on demand. On the other hand, a that they lack effective systems for realizing
substantial percentage coinsurance require- the benefits of new biologic therapies. On
ment—for example, 20%—can burden a one hand, the combination of growing con-
patient with annual costs well in excess of sumer power and less restrictive formularies
$2000. How then should health plans apply could translate into misuse of these costly
pharmacy management techniques to real- agents. On the other hand, aggressive con-
ize the twin goals of optimal patient health trols based solely on cost could deny
and sound financial management? patients’ treatments that slow progression
The RA market represents an interesting and delay longer-term medical costs.
case study. Etanercept and infliximab—the Like many conditions targeted by biologic
first tumor necrosis factor (TNF) antago- agents, RA presents a particularly challeng-
nists—have transformed RA treatment, and ing problem, given its comparatively low
created a market valued at more than $1 bil- incidence in managed care populations. Data
lion in 2000. Analysts project sales of TNF nevertheless suggest that the condition
antagonists to reach $4 billion by 2010. imposes a variety of costs, a point reinforced
Traditional agents have long been measured by advisers. In addition, the launches of
against relatively modest outcomes metrics, etanercept and infliximab fundamentally
metrics that may not capture the full bene- altered the economics of RA treatment,
fits of innovative biologic agents. increasing the costs of treatment and simul-
This raises a critical question: how best taneously challenging the existing cost-bene-
can the costs and benefits of biologic agents fit calculus. Although some believe that RA
in the context of a therapeutic area long char- represents a comparatively small component
acterized by disappointing outcomes and low of overall spending, the use of new biologic
expectations be evaluated? Indeed, if a new agents complicates the financial picture.
biologic agent offers the possibility of better Biologic therapies introduce not only dif-
outcomes and slowed progression, how can it ferent economics, but overall management
be evaluated effectively against existing clini- and administration challenges as well.
cal standards? Finally, with multiple biologic Pharmacy benefit managers generally esti-
agents to choose from, how should a health mate that injectable drugs are used by
plan manage patient access and choice? between 1% and 2% of the population yet
can account for up to 10% of drug costs.2
The Economic Challenge of Because many of the conditions treated with
Biologic Agents
As health plans shift costs to patients *The FDA approved adalimumab on December 31,
through tiered formularies and widening 2002.

S88 THE AMERICAN JOURNAL OF MANAGED CARE SEPTEMBER 2003

 Rheumatoid Arthritis Case Study

biologic agents occur so infrequently, oppor- optimal care represents a potentially expen-
tunities to negotiate volume-based price dis- sive policy.
counts are limited. In many cases, biologic At the same time, health plans must con-
agents and injectables are bundled with front their comparatively weak purchasing
treatment and administration costs at the power vis-à-vis manufacturers. Low inci-
provider level, and then passed on to payers dence has led to a situation in which special-
for reimbursement. This has the effect of ty practices (such as oncology, nephrology,
exposing payers to costs equivalent to aver- rheumatology, etc) now bundle their admin-
age wholesale price plus as much as 20% or istration and office visit costs with the cost
more, well in excess of the prices paid for of the biologic agent. Although this repre-
small molecule agents. Not surprisingly, sents an important revenue stream for
there is wide concern about provider-level providers, it has had the effect of exposing
incentives to overuse these agents. payers to prices well above average whole-
Current managed care systems remain ill sale price (AWP).
prepared to manage biologic agents and
injectables optimally. In fact, as MCOs
retreat from more restrictive plans—health
maintenance organization enrollment con- Although efficacy advantages will likely
tinues to decline, in favor of more open pre-
ferred provider organization and point of
be substantial, and may translate into
service plans—strategies to shift costs to tangible economic benefits, the managed
patients may place these therapies beyond care systems currently in place may be ill
the reach of most patients. Indeed, there is
already evidence that higher copay levels
suited for assessing and managing the
reduce patient demand, particularly for benefits of biologic agents.
higher-priced products.3

Applying Pharmacy Management Tools

to Biologic Agents. Herein lies the chal-
lenge: today the patient-as-consumer exerts Increasingly, payers have turned to spe-
considerable pressure on MCOs to eliminate cialty pharmacy vendors to help address this
restrictions and ensure nearly universal access problem. These vendors, which combine
to products, therapies, and diagnostics. In substantial volume-based purchasing power
exchange for this access, health plans, in with a variety of patient monitoring, over-
turn, extract a comparatively large share of sight, and disease management services,
product cost from the patient. That share offer payers prices below AWP, even when
increases for products deemed nonpreferred bundled with patient management. Further-
by the health plan. more, payers and specialty vendors have
Although this strategy may enable payer begun to push patient self-administration
organizations to drive patients toward less whenever possible, further limiting their
expensive therapies, it is by no means clear costs by eliminating reimbursements to
that it will maximize health outcomes. providers for office visits and administration.
Biologic agents frequently reside in the top Although the use of specialty pharmacy
tier of an MCO’s formulary, with copays services and patient self-administration will
ranging as high as 20% to 25% of the total both likely accelerate, neither adequately
cost. For many of the most promising bio- solves the principal dilemma facing payers
logic therapies, this out-of-pocket cost can at the dawn of the biologic era. Lowering
serve as a powerful disincentive to seek opti- acquisition and administration costs repre-
mal care. If, as clinical experience seems sents only short-term solutions, and does
increasingly to indicate, biologic DMARDs not address the broader question of how
slow disease progression, reduce costly hos- best to capture the often substantial clinical
pitalizations, and prevent joint replace- benefits of biotechnology while simultane-
ments, creating financial disincentives to ously managing its associated costs.

VOL. 9, NO. 4, SUP. THE AMERICAN JOURNAL OF MANAGED CARE S89

REPORT
Against this backdrop, biologic manufac- remains unclear, although there is broad
turers must define very specific value proposi- agreement that the majority of destruction
tions for their products while working to raise occurs in juxtaposition to the inflamed syn-
treatment expectations and refocus existing ovium, or pannus, that spreads to cover the
evaluative frameworks. Gaining formulary articular cartilage.4 Evidence shows that a
access with few restrictions will necessitate variety of cytokines, particularly inter-
a clear explanation of a product’s efficacy, leukin-1 (IL-1) and TNF alpha (α), play
tolerability, and compliance benefits, and a important roles.6,7 IL-1 and TNF-α stimulate
justification for the higher costs. A review of the cells of the pannus to produce collage-
RA treatment offers some key insights into nase and other proteases. They can also
this increasingly important challenge. stimulate the production of proteolytic
enzymes that degrade cartilage locally and
Strengths and Weaknesses of inhibit the synthesis of new matrix mole-
Small-Molecule Injectables cules. In addition, they may play a role in
RA is a chronic inflammatory auto- the activation of osteoclastic activity, result-
immune disease of unknown origin that ing in bone demineralization.
causes pain, swelling, stiffness, and ultimate- Recent advances in the understanding of
ly irreversible loss of function in the joints. cellular and molecular mechanisms of RA
Although RA has several special features that have highlighted the critical roles of IL-1
differentiate it from other kinds of arthritis, and TNF-α. The quest for chemically
the characteristic feature of RA is persistent amenable targets has recently led to the
inflammatory synovitis, usually involving identification and characterization of the
peripheral joints in symmetric distribution.4 intracellular signaling pathways associated
The disease may also affect other parts of the with these inflammatory cytokines. Because
body in addition to the joints. People with TNF-α has been found to play a central role
the disease often suffer from fatigue, occa- in the pathogenesis of the disease, it has
sional fever, and a general sense of malaise. proved to be a critical therapeutic target.6,7
The course of the disease varies signifi- There are indications that the biologic
cantly from person to person. Rarely, it lasts process of RA changes early in the disease,
only a few months or 1 to 2 years and goes so that patients may be less responsive to
away without causing any noticeable dam- treatment over time.8-10 One hypothesis
age. Other people have mild or moderate centers on the establishment of the pannus
disease, with periods of flares, and periods itself, suggesting that when the pannus has
in which they feel better, called remissions. formed, it becomes increasingly difficult to
Still others experience severe polyarthritic stall subsequent erosions.11 If, in fact, RA
disease that is active most of the time, lasts proceeds along a multistage pathway, late
for many years, and leads to serious joint intervention may miss opportunities to
damage and disability. Research shows that limit or slow progression. This question,
in some patients serious radiological damage which remains frustratingly unresolved, has
continues to progress more than 20 years important implications for the timing of
from disease onset.5 therapy initiation, the agent selected, and
the framework through which costs are
Pathophysiology and Natural History. evaluated.
RA is an immune-mediated disease charac-
terized primarily by dysregulation of T cell Current Treatment Strategies for
activity, which results in chronic, destruc- Rheumatoid Arthritis. Recent innovations
tive inflammation. The RA disease process in therapy have enhanced the quality of care
remains unclear, and the underlying etiolo- provided to RA patients. In addition, contin-
gy poorly understood, and there currently uing improvement in management has
exists no way to predict the progress from resulted in earlier diagnosis and treatment,
one inflammatory stage to the next. resulting in better patient outcomes and
The precise mechanism through which radiographic evidence of slowed or even halt-
bone and cartilage destruction occurs ed disease progression.

S90 THE AMERICAN JOURNAL OF MANAGED CARE SEPTEMBER 2003

 Rheumatoid Arthritis Case Study

The improvement in RA treatment never- clinical evidence suggests that earlier inter-
theless raises new, additional questions. For vention with a DMARD can slow progres-
example, if new pharmaceutical and biolog- sion, by reducing amplification or, later, by
ic agents offer dramatic gains in efficacy, interfering with the signaling associated
should expectations for remission be raised? with the lymphocyte recruitment process of
Is there sufficient evidence of improvement the second, chronic phase of inflamma-
beyond current American College of tion.14 Additional evidence reinforces that
Rheumatology (ACR) targets to justify the the best opportunity for slowing disease pro-
high price of biologic agents? When should gression occurs early in onset, possibly even
biologic intervention commence? These and before symptoms have become severe.8,9
other questions will only grow in impor-
tance as new products enter the market, and
treatment strategies evolve.
Ultimately, RA treatment seeks to slow or Whereas in the past, RA treatment sought
halt disease progression, preserving as much
patient function as possible. Until the 1980s, to reduce swelling and increase range of
the dominant treatment paradigm rested on motion, new agents make possible the
a series of sequential therapies that reserved maintenance of functional status and the
DMARD use until late in the course of the
disease.12 Although some patients did control of disease activity in a large num-
respond to minimal therapy, the majority ber of patients.
suffered poor outcomes, because of the
delay in adding disease-modifying agents to
the treatment regimen.
This approach gave way to a severity-
based strategy, still common today, using Evidence suggests that symptom scales
symptom-driven markers that included the and indexes of functional disability are less
number of swollen joints, loss of function, reliable indicators of disease progression than
persistence of pain, and elevated acute radiographic evidence, particularly during
phase reactants. Although it was an im- the earliest stages of the disease. The ACR
provement over the previous strategy, this guidelines and recommendations are set
approach resulted in poor outcomes as well, against the backdrop of research indicating
mainly because of the late use of disease- that joint damage progresses constantly dur-
modifying agents. ing the first 20 years after the onset of RA, as
Indeed, these approaches were predicated demonstrated by ongoing radiographic meas-
on 2 flawed assumptions: first, that DMARDs urement.20,21 Recent evidence appears to
should be used only later in the course of strengthen the relationship between joint
progression, after patients had failed on damage and disability in later (>8 years) stage
other agents, and second, that symptom- RA.20 The link between radiographic damage
based assessments were sufficiently sensitive and health assessment scale scores of disabil-
to the underlying joint deterioration charac- ity, although demonstrated to be very strong
terizing disease progression. in later stage RA, appear unrelated during
Disease duration has a considerable the earliest phases, reinforcing the impres-
impact on response to treatment, suggesting sion that significant, possibly irreversible
that the best opportunity for effective RA damage may take place before patient symp-
treatment occurs during the early stages of toms correlate with actual severity.
the disease.13 Multiple studies demonstrate
clear, clinical advantages of early, compara- End Points, Outcomes, and Treatment
tively more aggressive intervention with Options. Recent innovations in pharmaco-
DMARDs.14-19 logic therapy warrant a reexamination of
RA patients with longer disease duration treatment end points. Whereas in the past,
do not respond as well to treatment com- RA treatment sought to reduce swelling and
pared with patients with early disease,13 and increase range of motion, new agents make

VOL. 9, NO. 4, SUP. THE AMERICAN JOURNAL OF MANAGED CARE S91

REPORT
possible the maintenance of functional sta- tive to the true value of biologic agents. Yet
tus and the control of disease activity in a in the absence of an alternative evaluative
large number of patients. framework capable of providing a more
The ultimate goal of RA treatment is to robust picture of actual benefits, biologic
achieve complete remission. If remission is therapies are still evaluated against existing
not achieved, treatment seeks to control dis- benchmarks. At the same time, although
ease activity, alleviate pain, maintain func- clinical trial data from treatment-refractory
tion for activities of daily living and work, populations hint that even greater responses
and maximize QOL. The ACR’s 2002 RA are possible with earlier intervention, MCOs
management guidelines update recommends are unlikely to make approval decisions in
early intervention and initiation of DMARD the absence of strong data.
therapy within 3 months of an established Today, clinicians can choose from a vari-
diagnosis.22 ety of DMARDs, each offering a different
The ACR has developed criteria for defin- combination of efficacy, toxicity, price, and
ing improvement23 and clinical remission in durability. These include methotrexate,
RA, as follows24: antimalarials, leflunomide, sulfasalazine,
ACR criteria for 20% clinical improve- and biologic therapies. Agents should be
ment (the ACR20) require a 20% improve- evaluated based on a range of issues, includ-
ment in the tender and swollen joint count, ing safety, efficacy, toxicity, side effects, and
as well as a 20% improvement in 3 of the durability of clinical response. DMARDs
following 5 parameters: patient’s global have the potential to reduce or prevent joint
assessment, physician’s global assessment, damage, preserve joint integrity and func-
patient’s assessment of pain, degree of dis- tion, and, ultimately, reduce the total costs
ability, and level of acute phase reactant. of healthcare and maintain economic pro-
These criteria have been expanded to ductivity of the patient with RA.22
include criteria for 50% and 70% improve- Older DMARDs are characterized by their
ment measures (ACR50 and ACR70, relatively slow onset of action, unfavorable
respectively). toxicity profiles, and relative ineffectiveness
With new agents promising greater effica- in reducing erosions. Of greater concern is
cy, some researchers now argue in favor of the fact that there is little convincing evi-
raising the standards from 20% improve- dence that some of the older, premethotrex-
ment to 50% or even 70% improvement.25 In ate agents actually slow joint destruction,30
particular, the effectiveness of the TNF thus reducing their utility as anything other
agents suggests that a reconsideration of than palliative agents.
standard treatment guidelines is in order, Leflunomide is comparatively well toler-
particularly given the recognition that early ated, especially for patients that cannot tol-
diagnosis and early treatment are critical.26 erate methotrexate, although it has been
Indeed, it may now be possible to make sig- associated with gastrointestinal and liver
nificant strides beyond what have become problems. Randomized controlled trials of
standard therapeutic objectives. Some small leflunomide suggest that the agent slows
studies have shown promise against both radiographic progression.31 Although its
ACR5027 and in patients who had previously mechanism of action remains unclear,
failed on at least one DMARD.28,29 The possi- methotrexate has proved effective in slow-
bility of even greater clinical response with ing radiographic progression. It has proved
earlier use must be considered. to be an effective baseline agent to which
leflunomide and the newer biologic
Efficacy, Cost, and Expectations. Many agents can be added, providing additional
have expressed concern that, at present, the efficacy at lower doses of methotrexate.
biologic value proposition remains rela- It has, however, been associated with a high-
tively undefined and abstract. The current er incidence of hepatotoxicity, pulmonary
evaluative calculus, based on the use of non- complications, and myelosuppression.31,32
biologic DMARDs and relatively meager Selective targeting of pathogenic ele-
improvement goals, may, in fact, be insensi- ments of the disease has been shown to

S92 THE AMERICAN JOURNAL OF MANAGED CARE SEPTEMBER 2003

 Rheumatoid Arthritis Case Study

result in substantial improvement in signs duration do not respond as well to treatment

and symptoms, as well as disease progres- compared with patients who are in the early
sion.33,34 TNF-α antagonists, the first of these stages of the disease. Other predictors such
highly targeted products, have resulted in as being female, prior DMARD use, disease
rapid and substantial improvement in the functional class, and disease activity affect
signs and symptoms of disease as well as dis- response. This has implications for the clin-
ease modification, shown by slowing of radi- ical expectations of RA patients.37
ological progression. Obviously, the different types of DMARDs
Although further studies are needed to present important efficacy/toxicity trade-
clarify their ultimate position in the thera- offs,38 as well as significant differences in
peutic algorithm,7 the evidence of signifi- price and total cost. Older DMARDs can be
cant improvement is compelling. In one acquired relatively inexpensively, but offer
clinical trial, etanercept was shown to act less efficacy than the newer, more expensive
more rapidly than methotrexate to decrease agents, particularly the biologic agents.
symptoms and slow joint damage in patients Monitoring costs should also be considered,
with early active RA.35 Another study found particularly given the severity of some of the
that repeated doses of a TNF-α antagonist in known side effects.39 Similarly, because the
combination with methotrexate halted the use of the new biologic agents will likely
progression of joint damage over a 1-year increase treatment costs, particularly in the
period and resulted in improvement of radi- managed care setting, careful monitoring of
ographic scores of joint damage.36 Not- appropriate use may become more impor-
withstanding the efficacy of these agents, tant. It remains unclear what impact these
TNF-α plays an important role in host pro- strategies will have on monitoring costs.40
tection against infection and tumor genesis, Finally, some raise the subject of genetic
raising concern about serious infections and heterogeneity and its implications for treat-
malignancies. Data from trials, however, did ment effectiveness. Practical experience in
not show increased frequency of these the wake of the etanercept supply disrup-
events.7,22 tions suggests that the TNF antagonists
may have a disease-modulating effect. This
Considerations When Selecting a may suggest agent activity in changing the
DMARD Regimen. Although RA does not disease environment, possibly through gene
affect a large population, it presents unique involvement, although this is not clear.
treatment and management challenges Nevertheless, if this is a function of genetic
because of its complexity and severity. Even heterogeneity, exhibited in allele expres-
before biologic agents entered the market, sion, the issue bears further study, both in
providers had to weigh a variety of factors the context of RA and more broadly, for all
when selecting a treatment regimen. The biologic therapies.
introduction of new, substantially more
expensive agents—combined with steadily Economic Considerations in the
growing knowledge of the complex multi- Managed Care Setting. The annual direct
stage process of RA progression—has and indirect costs associated with RA in the
altered the decision-making framework. United States are estimated at, respectively,
As noted previously, evidence suggests $15 billion and nearly $50 billion.41
that the biologic process of RA changes over Estimated lifetime per-patient costs range
time, so that patients may become less from $61 000 to $122 000.41 As a whole,
responsive to treatment as the disease pro- musculoskeletal disorders impose a consid-
gresses. Because RA proceeds along a multi- erable burden upon society in terms of mor-
stage pathway, late intervention may miss bidity, long-term disability, and cost, but
opportunities to limit or slow progression, their impact in terms of mortality is low
most likely by failing to slow joint deteriora- compared with other diseases.42 At the same
tion before symptoms have become acute. time, many report difficulty in distinguishing
Treatment options must be considered in between different musculoskeletal condi-
this context. RA patients with longer disease tions when reviewing claims and related

VOL. 9, NO. 4, SUP. THE AMERICAN JOURNAL OF MANAGED CARE S93

REPORT
data, making an accurate plan-specific esti- Efficacy-Cost Evaluations for Biologic
mate of the burden more difficult. This prob- Agents in Managed Care Populations.
lem, which affects many of the low-incidence Early evidence with the biologic agents offers
conditions targeted by biologic agents, fur- encouragement. For example, data from
ther complicates economic evaluation. ongoing clinical work suggest that many
Estimates of the ratio of direct to indirect patients may be able to discontinue concur-
costs vary widely. Some studies estimate rent corticosteroid or methotrexate thera-
indirect costs to be more than double the py,47 thereby reducing polypharmacy costs.
level of direct costs.43 Other studies have Similarly, downward dose titration appears
found the ratio to be much closer to 1:1.39 feasible, an approach that should save costs
This ratio likely evolves as the disease pro- and prevent long-term side effects.48
gresses, and as patients experience decreas- Of potentially even greater interest are
ing function. cost-effectiveness data developed from
The majority of cost and economic stud- patients with methotrexate-refractory dis-
ies of the burden of RA conclude that ease. Preliminary findings appear to under-
inpatient costs represent the largest pro- score the importance of demonstrating
portion of direct costs, making less than economic value against a higher efficacy
one quarter of the RA population respon- standard.49 Indeed, what might the results
sible for at least 43%23 and up to 75%44 of have shown had ACR50 or ACR70 been
annual medical costs associated with RA. used as discrete outcomes measures?
Hospitalization costs were the highest com- Indirect costs studies have produced
ponent of direct costs, generally accounting mixed results regarding what point during
for 60% or more of costs whereas only the disease process that productivity losses
approximately 10% of patients with RA were begin in earnest. One study found that dis-
hospitalized. ability occurred in 25% of patients at 6.4
The only exception was a managed care years and in 50% at 20.9 years after disease
setting, in which hospitalization costs rep- onset, suggesting that most work disability
resented 16% of total direct costs. In man- occurs late in the course of the disease.50
aged care settings, costs of medications Research from the United States and
were proportionately higher than in fee-for- Finland found that 10 years after onset, 50%
service and inpatient settings,41 as a result of patients continued to work, and by 15
of differences in population demographics years this had decreased to between 33% to
and more aggressive attempts to offset med- 44%.51 Other studies appear to suggest the
ical costs with aggressive pharmaceutical opposite: Two studies from the Netherlands
therapy. found that the major loss of working capaci-
It remains unclear to what extent the ty takes place during the first 3 years after
care setting itself will affect DMARD use and onset.52,53
outcomes. Although earlier research com- Evidence shows that the direct medical
paring care settings45,46 found comparably costs associated with RA treatment rise
little difference, it did not include the newer dramatically in later stages of the illness; con-
biologic agents. Although MCOs generally versely, indirect costs, principally productivi-
exhibit lower hospitalization costs, their ty losses associated with disability, are
drug spending represents a much greater proportionally much greater in the early stages
percentage of total care, a function of both of the disease.54 Against this backdrop, delay-
younger, often healthier populations and ing the onset of work disability provides an
stronger economic incentives to offset med- economic rationale for earlier intervention:
ical costs with more aggressive upstream an individual in the United States who
interventions, including preventive care. As becomes disabled at the age of 40 can incur
a consequence, without a coherent eco- costs of several hundred thousand dollars in
nomic value proposition, one would expect indirect costs. These costs must be contrast-
that payer organizations may be more sen- ed with medical management costs.55
sitive to significant price increases in the The lack of agreement on this issue rais-
DMARD market. es important issues for both employers and

S94 THE AMERICAN JOURNAL OF MANAGED CARE SEPTEMBER 2003

 Rheumatoid Arthritis Case Study

their health plans. Work-related disability their old restrictive cost control models,
imposes significant costs on payers, particu- favoring instead “open” benefit designs that
larly early in the course of the disease, when pass the associated costs directly on to
patients remain relatively better able to con- patients.56 These designs, particularly mul-
tinue working. This has implications for titier formularies with graduated or per-
patients in managed care populations, par- centage copay differentials, have shown
ticularly given the epidemiology and demo- considerable promise in terms of sensitizing
graphics of the disease. patients to the economic consequences of
Not surprisingly, there are many con- their choices, particularly when selecting
cerns about productivity and disability- from a variety of similar products, such as
related costs. Although work-related proton pump inhibitors or statins.
disability represents the single largest socie-
tal burden, far surpassing total RA treatment
costs, participants believe that RA-related
costs are not easily explained to employers. Evidence shows that the direct medical
Although many of the more sophisticated costs associated with RA treatment rise
employers understand the value of a work-
er’s productivity, they are less clear about dramatically in later stages of the illness;
the impact of RA elsewhere in a house- conversely, indirect costs, principally
hold. If, for example, the employee serves productivity losses associated with dis-
as the primary caregiver to a family mem-
ber with RA, that has important implica- ability, are proportionally much greater
tions for productivity as well. in the early stages of the disease.
Lessons for Injectable Management
RA, by itself, represents a significant
challenge to MCOs. The combination of
comparatively low prevalence with high Benefit Management and Design. Many
costs and significant disability causes the continue to struggle with the benefit man-
development of effective management to agement issue. Although biologic agents for
become more difficult. Moreover, outcomes the treatment of RA target a comparatively
objectives have historically been unambi- small population, they represent significant
tious, reflecting both the rapid deterioration costs. The management of biologic thera-
of patient functional status and the absence pies, in fact, represents a significant chal-
of highly effective therapeutic options to lenge today and in the future, a challenge
forestall further erosions. In such a context, most MCOs have yet fully to address.
different products required a simple exami- For example, an important question for a
nation of the costs and benefits, with the payer is whether biologic therapies—for RA
understanding that patient deterioration and other conditions—are best classified as
was inevitable. medical or as pharmacy benefits. Classifying
The introduction of TNF-α inhibitors— biologic agents as a pharmacy benefit may
and of biologic agents in general—challenges shield the MCO from some costs (depending
this evaluative framework, by introducing on structure and design of the formulary)
the promise of substantial efficacy and toler- but may lead to a greater direct cost to the
ability improvements at a cost that com- patient via a higher copay, and may result
pletely overturns existing cost models. in lack of compliance, especially among
Complicating things further are issues such elderly or lower-income patients who have
as genetic predisposition to response, pro- limited pharmacy benefits. This will be par-
ductivity, and whether injectable or infusible ticularly true if MCOs institute a percentage
biologic therapies should be classified as copay, given the annual costs of biologic
medical or pharmacy benefits. agents exceed $12 000.
This all takes place at a time when payer As a medical benefit, on the other hand,
organizations continue to move away from the patient does not need to be concerned

VOL. 9, NO. 4, SUP. THE AMERICAN JOURNAL OF MANAGED CARE S95

REPORT
about the cost of individual biologic thera- intervention with DMARDs (although, at
pies, but the costs must then be absorbed by present, not specifically TNF agents) in
the payer organization. Given the physician- order to forestall joint deterioration even
level incentives to promote infused prod- before symptoms become acute.12 Yet few
ucts, the insurer may be challenged to track patients are satisfied with methotrexate—
utilization. Additionally, outpatient avail- apparently irrespective of its impact on dis-
ability may be constrained. ability, according to advisers. Acquisition
Greater patient access, paid for through costs would appear to justify a step-up
cost shifting, may not be compatible with approach to biologic agents; patient satisfac-
optimal RA treatment, particularly if tion, close to paramount today, seemingly
expensive DMARDs such as TNF-α receptor justifies the opposite.
modifiers are exposed to significant cost- Most believe that as newer biologic agents
sharing requirements. Patient willingness enter the market, a single agent could be
to pay may be insufficient to guarantee a assigned preferred status, albeit only in cer-
level of DMARD use sufficient to slow joint tain regulatory environments. This strategy,
deterioration, the hallmark of disease pro- like several other pharmacy management
gression, improve health-related QOL, or techniques, may be inappropriate for biolog-
reduce the direct and indirect costs associ- ic therapies, given the likely correlation
ated with the disease. Payer organizations between individual efficacy and genetic
may, in the course of limiting access to bio- specificity. In other words, a single agent
logic agents, forgo opportunities to reduce may show excellent results, but only in
medical costs and productivity losses by patients with particular TNF alleles. Giving a
slowing progression. It is possible, in fact, product advantages (or disadvantages) may
that some organizations may actually have negative effects, or unfairly penalize
increase their medical costs by driving some patients.
patients toward agents less effective against RA represents a complex and challenging
joint deterioration. condition for payer organizations. The
Nevertheless, in the current environ- introduction of new biotechnologies into the
ment, patient preferences and willingness to treatment armamentarium complicates
pay will increasingly play a role in the selec- matters still further. Because of their
tion of a therapeutic regimen, with impor- expense, biologic agents threaten to severe-
tant consequences. Advisers report that ly affect pharmacy budgets if managed
current levels of patient satisfaction with RA incorrectly. Yet the pharmacy management
drugs, whether methotrexate or biologic techniques currently in favor—tiered for-
therapies hover in the low teens (12% in 1 mularies, cost shifting, preferred products,
population). According to 1 adviser, before etc—are inappropriate to manage biologic
the introduction of the TNF agents, patients agents, particularly because of individual
remained on methotrexate for 5 to 8 years; genetic profiles. Moreover, narrowly concep-
today, average duration has declined to 14 tualized pharmacoeconomic evaluative
months.57 Although patients often seek the frameworks used to evaluate traditional
newest (and frequently the most expensive) DMARDs appear generally insensitive to
agents, these data suggest that patient satis- functional disability and patient QOL. If, as
faction issues will play a significant role in currently believed, biologic therapies offer
selecting therapy. substantial gains in these areas, existing
Data from Caremark (a leading prescrip- decision rules will no longer apply.
tion benefit manager) indicate that today, The era of consumer ascendancy and the
most RA patients start on a biologic agent, age of biologic agents have arrived simulta-
without first having failed on another agent, neously. As payer organizations try to rein-
and prescribing data suggest that rheuma- vent themselves, they must reevaluate their
tologists generate roughly 80% of the biolog- management strategies if patient satisfac-
ic agent prescriptions written for RA.58 tion and health outcomes are to be maxi-
Current clinical evidence supports early mized simultaneously.

S96 THE AMERICAN JOURNAL OF MANAGED CARE SEPTEMBER 2003

 Rheumatoid Arthritis Case Study

19. Plant MJ, Saklatvala J, Borg AA, et al. Measurement

REFERENCES and prediction of radiological progression in early
rheumatoid arthritis. J Rheumatol. 1994;21:1808-1813.
1. Weinblatt, ME, Keystone EC, Furst DE, et al. 20. Scott DL, Pugner K, Kaarela K, et al. The links
Adalimumab, a fully human anti-tumor necrosis factor between joint damage and disability in rheumatoid
alpha monoclonal antibody, for the treatment of arthritis. Rheumatology. 2000;39:122-132.
rheumatoid arthritis in patients taking concomitant
21. Janti JK, Kaarela K, Belt EA, et al. Incidence of
methotrexate: the ARMADA trial. Arthritis Rheum.
severe outcomes in rheumatoid arthritis during 20 years.
2003;48:35-45.
J Rheumatol. 2002;4:688-692.
2. Sirois P. HMOs should prepare now to get a handle
22. American College of Rheumatology Subcommittee
on biologics. Managed Care. 2002(April):42H-42N.
on Rheumatoid Arthritis Guidelines. Guidelines for the
3. Joyce GF, Escarce JJ, Solomon MD, Goldman DP. Management of Rheumatoid Arthritis: 2002 update.
Employer drug benefit plans and spending on prescrip- Arthritis Rheum. 2002;46:328-346.
tion drugs. JAMA. 2002;288:1733-1739.
23. Felson DT, Anderson JJ, Boers M, et al. The
4. Braunwald E, Fauci AS, Kasper DL, et al. Harrison’s American College of Rheumatology preliminary core set
Principles of Internal Medicine, 15th ed. New York: of disease activity measures for rheumatoid arthritis clin-
McGraw-Hill; 2001. ical trials. Arthritis Rheum. 1993;36:729-740.
5. Kaarela K, Kautiainen H. Continuous progression of 24. Meenan RF, Yelin EH, Henke CJ, et al. The costs of
radiological destruction in seropositive rheumatoid rheumatoid arthritis. A patient-oriented study of chronic
arthritis. J Rheumatol. 1997;24:1285-1287. disease costs. Arthritis Rheum. 1978;21:827-833.
6. Probert L, Plows D, Kontogeorgos G, et al. The type 25. Felson DT, Anderson JJ, Lange MLM, et al. Should
1 interleukin-1 receptor acts in series with tumor necro- improvement in rheumatoid arthritis clinical trials be
sis factor (TNF) to induce arthritis in TNF-transgenic defined as fifty percent or seventy percent improvement
mice. Eur J Immunol. 1995;25:1794-1797. in core set measures, rather than twenty percent?
7. Keystone EC. Tumor necrosis factor-α blockade in the Arthritis Rheum. 1998;41:1564-1570.
treatment of rheumatoid arthritis. Rheum Dis Clin North 26. O’Dell JR. How is it best to treat early rheumatoid
Am. 2001;27:427-443. arthritis patients? Best Pract Res Clin Rheumatol. 2001;
8. Emery P, Salmon M. Early rheumatoid arthritis: time 15:125-137.
to aim for remission? Ann Rheum Dis. 1995;54: 27. Weisman MH, Moreland LW, Furst DE, et al.
944-947. Efficacy, pharmacokinetic, and safety assessment of
9. Harris ED Jr. Rheumatoid arthritis: pathophysiology adalimumab, a fully human anti-tumor necrosis factor-
and implications for therapy. N Engl J Med. 1990;322: alpha monoclonal antibody, in adults with rheumatoid
1277-1289. arthritis receiving concomitant methotrexate: a pilot
10. Van der Heide A, Remme CA, Hofman DM, Jacobs study. Clin Ther. 2003;6:1700-1721.
JWG, Bijlsma JWJ. Prediction of progression of radio- 28. Temekomidis TI, Georgiadis AN, Almanos Y, et al.
logic damage in newly diagnosed rheumatoid arthritis. Infliximab treatment in combination with cyclosporine A
Arthritis Rheum. 1995;38:1466-1474. in patients with severe refractory rheumatoid arthritis.
11. Egsmose C, Lund B, Borg G, et al. Patients with Ann Rheum Dis. 2002;9:822-825.
rheumatoid arthritis benefit from early 2nd line therapy: 29. Farraccioli GF, Ascaloni R, DiPoi E, et al. Review
5 year follow-up of a prospective double blind placebo of combination therapy failures using infliximab while
controlled study. J Rheumatol. 1995;22:2208-2213. maintaining the combination or monotherapy with
12. Wolfe F, Rehaman Q, Lane NE, et al. Starting a dis- methotrexate: results of an open trial. Rheumatology.
ease modifying antirheumatic drug or a biologic agent in 2002;41:1109-1112.
rheumatoid arthritis: standards of practice for RA treat- 30. Iannuzzi L, Dawson N, Zein N, et al. Does drug
ment. J Rheumatol. 2001;28:1704-1711. therapy slow radiographic deterioration in rheumatoid
13. Anderson JJ, Wells G, Werhoeven AC. Factors pre- arthritis? N Engl J Med. 1983;309:1023-1028.
dicting response to treatment in rheumatoid arthritis: the 31. Sharp JT, Strand V, Leung H, et al. Treatment with
importance of disease duration. Arthritis Rheum. leflunomide slows radiographic progression of rheuma-
2000;43:22-29. toid arthritis: results from three randomized controlled
14. Perlman H, Pagliari LJ, Volin MJ. Regulation of trials of leflunomide in patients with active rheumatoid
apoptosis and cell cycle activity in rheumatoid arthritis. arthritis. Leflunomide Rheumatoid Arthritis Investigators
Curr Mol Med. 2001 Nov;1:597-608. Group. Arthritis Rheum. 2000;43:495-505.
15. Irvine S, Munro R, Porter D. Early referral, diagno- 32. Capell HA. Disease modifying antirheumatic drugs:
sis, and treatment of rheumatoid arthritis: evidence for longterm safety issues. J Rheumatol. 2001;62(suppl):10-15.
changing medical practice. Ann Rheum Dis. 1999;58: 33. Haraoui B, Strand V, Keystone E. Biologic agents in
510-513. the treatment of rheumatoid arthritis. Curr Pharm
16. Wolfe F, Albert DA, Pincus T. A survey of United Biotechnol. 2000;1:217-233.
States rheumatologists concerning effectiveness of dis- 34. Gabriel SE, Coyle D, Moreland LW. A clinical and
ease-modifying antirheumatic drugs and prednisone in economic review of disease-modifying antirheumatic
the treatment of rheumatic arthritis. Arthritis Care Res. drugs. Pharmacoeconomics. 2001;19:715-728.
1998;11:375-381. 35. Bathon JM, Martin RW, Fleischmann RM, et al. A
17. Boers M. Combination therapy in rheumatoid arthri- comparison of etanercept and methotrexate in patients
tis. Lancet. 1999;354:952. with early rheumatoid arthritis. N Engl J Med. 2000;
18. Mottonen TT, Hannonen PJ, Boers M. Combination 343:1586-1593.
DMARD therapy including corticosteroids in early rheuma- 36. Lipsky PE, Van der Heide MFM, St. Clair EW, et al,
toid arthritis. Clin Exp Rheumatol. 1999;17:S59-S65. for the Tumor Necrosis Factor Trial in Rheumatoid

VOL. 9, NO. 4, SUP. THE AMERICAN JOURNAL OF MANAGED CARE S97

REPORT
Arthritis with Concomitant Therapy Group. Infliximab 48. Den Broeder AA, Creemers MC, van Gestel AM,
and methotrexate in the treatment of rheumatoid arthri- van Riel PL. Dose titration using the Disease Activity
tis. N Engl J Med. 2000;343:1594-1602. Score (DAS28) in rheumatoid arthritis patients treated
37. Cooper NJ. Economic burden of rheumatoid arthri- with anti-TNF-alpha. Rheumatology. 2002;41:638-642.
tis: a systematic review. Rheumatology. 1999;39:28-33. 49. Taylor PC. Anti-TNF-alpha therapy for rheumatoid
38. Felson DT, Anderson JJ, Meenan RF. Use of short- arthritis: an update. Intern Med. 2003;1:15-20.
term efficacy/toxicity tradeoffs to select second-line 50. Wolfe F, Hawley DJ. The long-term outcomes of
drugs in rheumatoid arthritis. Arthritis Rheum. 1992;35: rheumatoid arthritis: work disability: a prospective 18 year
1117-1125. study of 823 patients. J Rheumatol. 1997;25:2108-2117.
39. Comer M, Scott DL, Doyle DV, et al. Are slow-act- 51. Wolfe F. 50 years of antirheumatic therapy: the
ing anti-rheumatic drugs monitored too often? An audit prognosis of antirheumatic arthritis. J Rheumatol.
of current clinical practice. Br J Rheumatol. 1995;34: 1990;17(suppl 22):24-32.
966-970. 52. Van Jaarsveld CHM, Jacobs JW, Schrijvers AJ, et al.
40. Lubeck DP. A review of the direct costs of rheuma- Effects of rheumatoid arthritis on employment and social
toid arthritis: managed care versus fee-for-service set- participation during the first years of disease in the
tings. Pharmacoeconomics. 2001;19:811-818. Netherlands. Br J Rheumatol. 1998;37:848-853.
41. Yelin E, Callahan LF. The economic cost and social 53. Van Jaarsveld CHM. The Utrecht Rheumatoid
psychological impact of musculoskeletal conditions. Arthritis Cohort: on treatment, prediction, and cost
Arthritis Rheum. 1995;38:1351-1362. [thesis]. Utrecht. Utrecht University, 1998.
42. Rice DP, Hodgson TA, Kopstein AN. The economic 54. Pugner KM, Scott DI, Holmes JW, et al. The costs
costs of illness: replication and update. Health Care of rheumatoid arthritis: an international long-term view.
Financing Review. 1985;7:61-80. Semin Arthritis Rheum. 2000;29:305-320.
43. Pugner K, Scott D, Holmes J, et al. The costs of 55. Pincus T, Mitchell JM, Burkhauser RV. Substantial
rheumatoid arthritis: an international long-term view. work disability and earnings losses in individuals
Semin Arthritis Rheum. 2000;29:305-320. less than age 65 with osteoarthritis: comparisons with
44. Lanes SF, Lanza LL, Radensky PW, et al. Resource rheumatoid arthritis. J Clin Epidemiol. 1989;42:
utilization and cost of care for rheumatoid arthritis and 449-457.
osteoarthritis in a managed care setting—the importance 56. Robinson JC. Renewed emphasis on consumer cost
of drug and surgery costs. Arthritis Rheum. 1997;40: sharing in health insurance benefit design. Health Aff.
1475-1481. 2002. Available at: http://www.healthaffairs.org/
45. Yelin E, Trupin L, Katz P. The prevalence and impact WebExclusives/Robinson_Web_Excl_032002.htm.
of managed care for persons with rheumatoid arthritis in Accessed June 10, 2002.
1994 and 1999. Arthritis Rheum. 2002;47:172-180. 57. Wolfe F, Michaud K, Stephenson B, et al. Toward a
46. Ward MM, Lubeck D, Leigh JP. Long-term health definition and method of assessment of treatment failure
outcomes of patients with rheumatoid arthritis treated and treatment effectiveness: the case of leflunomide ver-
in managed care and fee-for-service practice settings. sus methotrexate. J Rheumatol. In press.
J Rheumatol. 1998;25:641-649. 58. Berger J, Fensterheim L, Slezak J, et al. Appropriate
47. Phillips K, Husni ME, Karlson EW, Coblyn JS. use of TNF therapy for rheumatoid arthritis. Poster pres-
Experience with etanercept in an academic medical entation. American College of Rheumatology Annual
center: are infection rates increased? Arthritis Rheum. Meeting. October 27, 2002.
2002;47:1-4.

S98 THE AMERICAN JOURNAL OF MANAGED CARE SEPTEMBER 2003