Você está na página 1de 10


Escaping the sword of Damocles:

Toward a new future for
pharmaceutical R&D
McKinsey perspectives on drug and device R&D 2012 3
Escaping the sword of Damocles: Toward a new future for pharmaceutical R&D

Recent years have seen a collapse in the industrys R&D productivity

and a loss of faith in its innovation model. Regaining customers and
shareholders trust by delivering life-changing new drugs is still an
achievable goal, but it will require discipline, creativity, and luck in
equal measure.

Ajay Dhankhar, Matthias Evers, and Martin Mller

Damocles was a courtier in Greece in A decade of doubt

the fourth century BC. The story has
it that he used to flatter the king by The pharmaceutical and biotech industry
saying what a marvelous life he had. has failed to meet shareholder expectations
When the king offered to swap places over the past decade, and has come
with him for a day, Damocles agreed, nowhere near beating the R&D odds.
only to find himself sitting beneath Indeed, R&D looks like a rigged game.
a huge sword that was hanging by Though a few companies have bucked the
a single hair from a horses tail. He trend, the jury is still out on whether they
couldnt move without putting his are making genuine improvements to their
life in danger. The episode taught models that will stand the test of time.
Damocles a sharp lesson about the
gravity of a leaders responsibilities. In the past 25 years the industry has
created in excess of $1 trillion of
The trends of the past few years can be shareholder value, but destroyed around
likened to a sword of Damocles hanging $550 billion of value during the decade
over the pharmaceutical industry. Yet of doubt from 2000 to 2010. That value
there are good reasons for continuing destruction coincided with a 60 percent
to believe in it. Unmet needs, scientific increase in the R&D spending rate from
advances, and increasing affluence should 10 to 16 percent of sales, and with an
translate into continuing opportunities to even higher increase in absolute spend
innovate for the benefit of patients. We as worldwide sales grew from $200
expect to see evolution at the core and billion in 1995 to $800 billion in 2009.
revolutions at the periphery, as well as
some fundamentally new R&D ideas. A recent McKinsey analysis calculates that
the average economic return on R&D has
So what does this mean in practical dropped from between 13 and 15 percent
terms? As we discuss below, companies in the 1990s to between 4 and 9 percent in
must adopt a different approach to the past decade (Exhibit 1). This suggests
their R&D spend, create more exciting that much of the current investment in
environments to attract the brightest R&D is not creating value. We estimate
scientists, find ways of creating an that cumulative success rates have fallen
ownership mindset, and embrace by as much as 50 percent as the number
collaboration and co-invention to take R&D of drug development programs and the
beyond the walls of their organization. cost per program have doubled.1 For the

companies under the most Exhibit 1: Economic return on R&D investment

pressure, the net present value for top 10 biopharma players
Includes impact of working capital, property, plant, and equipment, and goodwill
of their pipeline is negative.
Forecast Performance
returns trajectory
Not surprisingly, stakeholders 1314%

and shareholders are losing 1012%

patience and exerting 68%

mounting pressure on 45%

boards, CEOs, and executive

teams to acknowledge the 199195 19962000 200105 200610 Current Step-change
trajectory (30% further
situation and reduce R&D productivity
costs. In addition, it is widely
believed that one-off launches Percent of sales
reinvested in R&D
910% 12% 16% 17% 1516% ?

may show only ephemeral Decade of Decade of high The way

improvements in return on high returns re-investment forward?

investment and encourage

bravado, hiding deeper issues
about growing trial costs, Exhibit 2: Players are losing their ability to
falling success rates in virtually outperform the market
Selection of top 20 pharma and biotech players
all therapeutic areas (TAs) Total return to shareholders, CAGR, percent

and molecule types, more 19852010 20002010

crowded markets, higher bars Company A 23 -2
B 18 18
for commercial success, and C 15 4

the unexpectedly swift loss D

of the partnering advantage. F 13 3
G 13 -5
H 12 -6
I 11 -1
As yet there is no evidence J 10 -1
that the trend has bottomed K
L 9
10 -6
out and success rates are M 9 -5
N 7 -3
improving. Things may get O 7 -1

worse before they get better, S&P 500 S&P 500

index: 9.9 index: 1.4
a view endorsed by most
serious industry observers.

Admittedly, some companies

have beaten the odds, but whether whether TA specialization is a winning
their success is down to sustainable model, or whether the future might lie in
value creation or serendipity is unclear. higher exposure to biologics, for every
Many pharmaceutical companies have such trend there are counter-examples
had significant 25-year shareholder and reasons to suppose that the opposite
value creation, although their results conclusion might be equally valid.
for the past ten years are more modest
(Exhibit2). These success stories dont
point to one promising direction that the
industry can follow; rather, several fields
have pockets of excellence that seem to
pay off. Tempting though it is to wonder
McKinsey perspectives on drug and device R&D 2012 5
Escaping the sword of Damocles: Toward a new future for pharmaceutical R&D

The environment is getting tougher on drug profiles, and the

getting tougher US is no longer immune. As real-world
outcomes become more and more
Those who take a pessimistic view important, there is limited willingness to pay
can point to still more headwinds that for efficacy alone. Countries with formal
will hold back R&D productivity. cost-effectiveness assessments in drug-
funding decisions now account for some
Most low-hanging fruit has already 60 percent of global prescription sales, a
been picked. Libraries have been number that is growing fast. As a result,
screened and monoclonal antibody most companies internal innovation hurdle
approaches have been run on all has shifted beyond me too strategies
obvious extra-cellular targets. Expensive toward earlier screening (as early as lead
technology investments in such areas optimization) for differentiation against
as functional genomics have not yet the evolving standard of care. As payors
paid off, and it is unclear whether they grow ever more sophisticated and more
ever will. The industry is suffering from and more technologies and techniques
a surfeit of similarity, as evident in the for personalized or protocolized
massive competition in oncology and healthcare become available, the
elsewhere among many players circling differentiation requirements for individual
a handful of targets. No one has really drugs will become increasingly specific.
cracked how to capture advantage from
the emerging science around disease Returns for many companies will
biology and understanding, biomarkers, deteriorate further. That isnt because
and model-based drug development. there are no advances left to make, but
because too many duplicative bets are
Regulatory environments remain being placed by relatively low-skilled
challenging in the post-Vioxx world. resources that are the legacy of excess
New medicines are unlikely to be approved investment during the artificially high profit
without major risk management measures umbrella of the late 1990s. Put simply,
or label restrictions. The progress made this is a case of overcapacityand the
by regulatory science in adapting to capacity with the lowest productivity
new model-based drug development will be removed from the market. This
approaches has been limited. Recent is already happening through the R&D
favorable reviews of applications appear to restructurings, mergers and acquisitions,
reflect good science rather than a change and site closures seen throughout the
in processes, productivity, or risk tolerance. industry in the past couple of years.

Remnants of the old shots on goal

paradigm persist in the portfolio.
High attrition in Phase II and III may Not all doom and gloom
continue for several more years if lower-
quality compounds continue to be pushed For the optimists among us, however, there
forward instead of getting weeded out. are bright spots that provide some hope.

A major new post-approval hurdle Investigational new drug (IND) filings

has emerged. Pricing, reimbursement, have come down by 17 percent in the
and health technology assessments are past few years. This is a clear sign that

excess and unproductive Exhibit 3: IND filings decline

Number of commercial investigational new drugs
capacity is starting to be
removed (Exhibit 3).
17% pa
Numerous players +5% pa 883
are piloting new ideas 779
successfully. Examples
include Novartiss pathway
approach; multiple companies Migration away
from shots on
proof-of-concept strategies; goal model

heavyweight teams and

streamlined decision-making
1986 1990 1995 2000 2005 2010
processes; GlaxoSmithKlines
modularization into ever- * Estimate

Data after 2004 includes INDs for therapeutic biologic products transferred from CBER to CDER
smaller performance units; Source: CDER; FDA
Lillys Chorus; numerous
Covance-like contractresearch
organization(CRO)deals;and Regulators are starting to recognize
many partnerships. that regulatory science must
improve. They are also beginning
The industrys understanding of biology to understand that a new type of
is expected to improve over the next dialogue with industry is needed.
decade. Entrants with new talents, skills,
and orthogonal perspectives are joining Electronic health information (EHI),
the party: the NIH, the FDA, academia, e-trials, and real-world evidence could
the Bill & Melinda Gates Foundation, and create significant value across the
many governments. Fresh opportunities product lifecycle. For example, they
may emerge in modeling and simulations, could inform trial design and decision
biomarker identification and usage, making and improve market access by
and the use of outcome data as a way providing more robust data on comparative
to focus and guide clinical trials. The effectiveness and safety (Exhibit 4).
potential opportunity, and big cost, of
massive bioinformatic and genomics,
proteomics, and metabolomics tools and
insights could finally start to pay off. Evolution at the core,
revolutions on the periphery
These advances could eventually open
the door to the world of personalized The R&D strategy and operating model
healthcare. This would present major we see for the future is one forged
uncertainties for the industrys business around variablizedand in most but not
model, but clear opportunities for better all cases reducedspend. We also see
treatment of individual patientsand hence evolutionary but deep changes at the core,
commercial potential. Better biology, better complemented by targeted revolutionary
and less costly genomics, and personalized bets in a few game-changing areas.
medicine may also allow some failed This will require an overall reduction in
molecules of the past to be resurrected. the number of programs, a Darwinian
discipline in portfolio development and
McKinsey perspectives on drug and device R&D 2012 7
Escaping the sword of Damocles: Toward a new future for pharmaceutical R&D

Exhibit 4: How EHI can add value Economic


Development Approval and launch Growth and maintenance



New indicators/ Loss of exclusivity/

Phase III Approval/HTA New competition
formulations competition
Define ideal patient Assess burden of Support continued Demonstrate Differentiate from
population/unmet disease/unmet success unmet need competitors going
need need Additional Identify target generic
Demonstrate Estimate cost evidence on populations Differences in usage
Uses of burden of disease effectiveness and clinical outcomes Outcomes in sub-
real-world Recruit trial budget impacts Differentiation populations
evidence patients accurately (vs. usage and Cost effectiveness
Understand standard of care) prescribing Effects of switching on
standard of care Meet post- patterns outcomes
and clinical and marketing Adherence to Adherence to drug
cost effectiveness requirements drug regimen regimen
to achieve a (safety, utilization) Differentiate with or
reimbursable against protected
product profile galenics

decision making so that only the strongest Variablize and possibly

programs survive, and an ownership reduce R&Dinvestment
mindset among R&D leaders and project The days of the shots on goal model
teams so that resources are used much are numbered. There are not enough
more thoughtfully, as if we owned the quality pipeline assets and validated
assets and the company ourselves. targets in discovery or the clinic to launch
so many shots while maintaining a
We expect companies to focus on formulaic investment of 15 to 20 percent
well-known levers to make the smaller of sales. Instead, we expect companies
number of programs more effective. to take quality shots on goal starting
Reorganizations and mergers will be from new libraries and sources of targets.
much less important than, for example, Standard high-throughput screening (HTS)
quality of governance, senior team approaches and numbers-based incentives
decision-making processes, metrics, will be supplemented or even abandoned.
incentives, and a culture of innovation.
We also expect to see some creativity Its time to make the level of R&D
and willingness to experiment. spending more flexible. R&D outlay
need not be fixed at 15 percent of
Our view of what will drive superior R&D revenue, nor at the 1990s level of
productivity is based on lessons from 10percent. Instead, companies could flex
the past as well as the pressures and it between 5 and 25 percent depending
opportunities we have outlined. Some of on portfolio quality, pipeline evolution,
our predictions are well supported and and fluctuations in the quality of external
consistent with industry views; others assets. They could pursue opportunities
are more speculative and controversial. that show genuine promise and be ready

to reduce or increase funding as each Enhance the environment you offer.

case dictates. Before they can do this, Make your R&D organization the Apple or
though, companies will have to dismantle Google for ambitious scientists. Attracting,
fixed infrastructurea process that has developing, and ensuring collaboration
already started across the industry. among the brightest researchers and
drug hunters truly matters.2 Place as
Redundant capacity must go. Obvious much emphasis on creating a stimulating
overlaps are already being removed environment as on driving efficiency.
through partnerships in R&D, such as
that between Boehringer Ingelheim Ensure clear differentiation in a
and Eli Lilly in diabetes. Partnerships challenging payor environment. This
and alliances are a natural way to is about medical and clinical and cost-
reduce capacity while continuing to effective differentiation, not just novelty.
access good science in the therapeutic Creating cross-functional alignment
areas that are strategically valuable. on what differentiation means and
allocating funds appropriately are key.
Teams should act as owners, not So is conducting evidence-based drug
managers, of R&D assets. The concept development in real-world settings.
of better owner has been poorly applied
to R&D assets. It requires a mindset Make the most of your differentiated
that an R&D team doesnt consider assets. Improve the effectiveness of
itself distinctive unless it genuinely is, your lifecycle management (LCM) as
and leaders who are prepared to make a way to add value to a franchise. The
dispassionate decisions to sell or licence scarcest and hence most valuable of
out compounds that may be more all assets is an approved molecule. It is
valuable in others hands. For example, it important to create a franchise that can
is not clear that many companies can be expand the brand, perhaps even beyond
distinctive in more than five therapeutic the active pharmaceutical ingredient,
areas and multiple disease biology areas while maintaining the brand equity.
unless they have huge budgets and scale.
Better ownership also requires leaders who Take a Darwinian approach to decision
view investments as if they were their own, making. Evaluating the portfolio objectively,
and companies that enable and empower eliminating decision-making biases,
them to do so. Companies should create and allowing only the best programs to
incentives to kill programs when necessary, survive are critical. We find its almost
and make it clear they do not regard a impossible for a management team
program kill as a career-limiting move. of non-scientists to act as responsible
stewards of a research portfolio;
Pursue evolutionary but deep conversely, scientific teams often find it
changesat the core difficult to be dispassionate. Companies
R&D will not be transformed overnight, seldom get a truly independent read on
nor will there be a paradigm shift. their pipeline quality, but when they do,
The priority should be purposeful it can yield valuable insights. Possible
execution against well-known but approaches to achieve this include creating
often poorly executed levers: a blue-ribbon FDA that applies the
same level of scrutiny to a draft dossier
as the FDA would, bringing the same
McKinsey perspectives on drug and device R&D 2012 9
Escaping the sword of Damocles: Toward a new future for pharmaceutical R&D

Exhibit 5: Segmenting the portfolio into swim streams Devise a new incentive
Based on estimates of approximate aggregate attrition rates for medicines in the following therapeutic areas:
central nervous system, endocrine, cardiovascular, infectious diseases, oncology, and respiratory model. Basing incentives and
Quadrant D Quadrant C goals only on the number of
FAST LANE filings or the size of a portfolio
More destroys more value than
perhaps any other action in
Attrition rate 60% Attrition rate 25%
of endpoints*
the industry. To rekindle a
Attrition rate 70% Attrition rate 35%
culture of innovation while
Less simultaneously managing
scientists, leaders need to
Quadrant A Quadrant B
create performance metrics
No Yes
and incentives that promote
Established mechanism of action

R&D quality and output

* More objective endpoints relate to more easily reproducible diagnostic tests or measures, as opposed

to less reproducible scales or patients self-reporting diaries
Novelty of mechanism is more relevant than objectivity of endpoints
rather than just throughput
Source: Evaluate; Pharmaprojects; Factiva; McKinsey analysis efficiency (which often takes
care of itself when resources
are constrained). Companies
cross-functional lens to evaluate internal should put in place a system that enables
assets as in-licenced molecules, and the best biologists and chemists to work
adopting a venture capitalists approach in the highest-value areas and allows
to R&D decisions. Indeed, the trend them to have portfolios at all levels, an
toward more VC and investor funding excess of ideas and investment options,
of development programs may well be and limited funds. Instead of putting
driven by the dispassionate analysis that people in a position where they have
such leaders bring to decision making to prosecute bad molecules to avoid
rather than by the funding itself, which ending their careers, give them incentives
usually comes at a high cost of capital. to suggest better avenues to pursue.

Avoid making Toyotas in a Lexus Improve basic efficiency and

factory. Companies should consider effectiveness. High levels of waste and
segmenting their portfolio into swim gold-plated solutions can still be found in
streams that move at different speeds R&D, and indeed in pharma as a whole.
through steady waters or rapids, Staff who join from other industries
internally and externally (Exhibit 5). They are frequently surprised by the lack of
should systematically differentiate the discipline in cost management. Companies
way they treat R&D projects not just by should adopt methods such as lean,
value, but also by risk and data clarity. outsourcing and offshoring, and external
This would determine how teams are spend management and oversight.
staffed, how much frontloading to do,
and when it is necessary to go external. Amplify your discovery and clinical
Companies should also decide their research expertise. It is extraordinarily
strategy in terms of which water to challenging to design laboratory or clinical
swim inthe kiddie pool or the piranha- experiments that are both informative
infested stream?for each therapeutic under all possible outcomes and tailored to
area and for the portfolio as a whole. regulatory and real-world success factors.
Too many experiments fail because of
subtle design flaws. Developing a pool

of seasoned researchers is one of the discussion on the optimal size of an R&D

most obvious productivity levers, yet unit. Is it 200 to 300 researchers or as few
many get it wrong. Every company has a as 50 to 70? Or should even smaller units
small group of world-class researchers; coordinate networks of increasingly global
the best companies figure out how to contract research organizations (CROs)
amplify their contributions by helping to get the work done, while planning,
them build the next generation of leaders strategy, and design are the preserve of a
in scientific and medical research. team of high-caliber scientists and medics?
More companies are likely to experiment
Consider revolutions at the periphery with such models. In time, they may even
Potential game changers or new lead to the complete disaggregation of
paradigm solutions include: the industry value chain as CROs take
over the lions share of operational work.
Next-generation licencing or drug co-
invention. If pharmaceutical companies
could collaborate as effectively as high-
tech and movie companies do, significant Revisiting R&D strategy
value could be created. Biology research
should happen less through in-house Although it would be unwise to over-
efforts and more through early-stage generalize about R&D operating models,
collaborations. Strategy should revolve our outside-in view suggests that most
around fractional bets on a larger portfolio companies have room to improve. They
of molecules. Opportunities exist to dont have to nail every single factor that
separate out who funds, who prosecutes, we have highlighted, but they do need a
and who markets a molecule, and to base level of performance in most of them,
craft multi-party agreements to make coupled with genuine distinctiveness in a
that happen. Another way to create a few. Most companies would find it useful
co-invention ecosystem is to undertake to consider the following questions:
deep collaborations with academics.
Instead of setting a top-down budget,
such as dedicating 15 percent of
A scale-up of faster, cheaper drug
investment to R&D, should we assess
to proof of concept paradigms.
our pipeline and external options
If the Chorus model proves to be
as candidates for investment and
feasible at scale, it could be emulated
build a bottom-up budget to allow
by others. Pharmaceutical companies
greater flexibility from year to year?
could do what carmakers do and work
with multiple partners in emerging What are the therapeutic and other
markets to help them develop from areas where we are truly distinctive
service providers with individual slivers and have critical mass? Would a
of the value chain to more integrated venture capitalist or the FDA reach
participants in the development process. the same conclusions? Should
we refine the number and mix of
Small, empowered, entrepreneurial therapeutic areas we cover?
R&D units. Ever since GlaxoSmithKline Could we embrace and institutionalize a
launched its Centers of Excellence for mindset to address the fourth hurdle
Drug Discovery (CEDDs) concept more to developmentthe market access
than 10 years ago, there has been much challengeto ensure effective LCM?
McKinsey perspectives on drug and device R&D 2012 11
Escaping the sword of Damocles: Toward a new future for pharmaceutical R&D

How Darwinian are our R&D

governance and decision-making
processes? Are there biases we
After a decade-long crisis in R&D
should eliminate? Do we strike the
productivity, there is much sound thinking
right balance of risk for internal
on how to do things better. Whats
and external candidates?
more, many companies are improving
What could we do to improve our parts of their business, and some are
efficiency and effectiveness? managing to outperform in most or all
How could we benefit from of it. The real challenge is being able
broader partnerships, drug co- to change at scale: not only individual
invention approaches, and an functions and therapeutic areas, but major
environment of borderless R&D? companies and ultimately the industry
What other revolutions could as a whole. Perhaps pharma will then be
we embrace: faster drug to able to put its decade of doubt behind
proof of concept paradigms, it and embrace a decade of change.
more entrepreneurial R&D units,
government collaborations?

Companies have tried or are trying

most if not all of the approaches we
have described above. It isnt yet clear
what will work and what wont. The
right mix of interventions is likely to vary
from one company to another, given
the differences in starting points.

1 For more detail on the decline in success rates, see The anatomy of attrition revisited, pp. 247.
2 For more on this topic, see Managing the health of early-stage discovery, pp. 2833.

Ajay Dhankhar is a director in McKinseys New Jersey office, Matthias Evers is a principal in the Hamburg
office, and Martin Mller is a principal in the Copenhagen office. The authors wish to acknowledge the
contributions of many colleagues to this article, in particular Lynn Dorsey Bleil, Sylvain Milet, Lucy Perez,
Tejash Shah, Nav Singh, and Nicole Szlezak.