| |
PDF version
|
|
|
R&D
planning and prioritisation
Any self-respecting pharmaceutical R&D unit is constantly
coming up with new ideas for products that could be developed
– most of them described by their originators as potential
blockbusters – and the rate at which these are being
brought forward is beginning to increase dramatically with
the introduction of more efficient discovery processes such
as high-throughout screening. At the same time the corporate
Business Development function is bringing in a steady supply
of products ideas from third parties which could be added
to the list of internal development projects.
The decision on which of these should be developed in-house,
which licensed out or otherwise disposed of and which should
just be abandoned is one of the biggest that any pharmaceutical
company must make. It is crucial to identify at an early stage
which products might be able to provide the engine of growth
into the future – and equally important to weed out
those that will be a hugely expensive waste of time and effort.
In the past the industry has not had a good record of making
the correct decisions. Together with some colleagues the author
recently analysed the performance of all products introduced
on to the pharmaceutical market in the years 1991 and 1992,
to find out whether how many of them had been worth the massive
development effort put into them. We excluded products launched
only in Japan, and took into account all presentations and
indications for a given compound. The result does not make
good reading. A small number of the products did very well
– the group includes the compounds paroxetine, lansoprazole
and sumatriptan, which achieved peak sales in excess of $1
billion. But a disturbingly large number achieved virtually
nothing - nearly 40% of them had estimated peak sales of less
than $25 million, and many virtually zero, as the chart below
shows:
Considering that companies currently are forecasting an output
rate from R&D of only about one product for every $1 billion
spent, it is obvious that this is a situation that must be
avoided – or at least substantially improved - in the
future. The way to do it is by better portfolio selection
processes in the early stages of R&D.
This paper describes a methodology which has been shown to
work in such situations, and which avoids complex financial
calculations while seeking to minimise the complications caused
by the many differences of opinion and behaviour of those
involved in the decision-making process. The main steps are:
Defining what the company is trying to
achieve with its product portfolio Breaking down this strategy into component targets against
which projects can be measured Assessing the projects and their markets to provide information
which will permit them to be scored against each these targets Scoring the projects and assessing the results.
Defining what the
company wishes to achieve
Any discussion about the possible adoption of a new project
needs to be based on a clear understanding of what the company
concerned is trying to achieve. Presumably the project would
not have been proposed for adoption if it did not have at
least some potential to succeed commercially. But how much
commercial success must it have? When must it achieve this
target? Will it support and play to the company’s strengths
in the market-place, or will it lead in a completely new direction?
Will it require new investment in skills and resources in
R&D, or can it be put through existing machinery? How
likely is it that the product will complete development successfully
at all, and how happy is the company with the level of risk
that it might fail in development?
It is remarkable in the author’s experience how rarely
those involved in making or supporting portfolio decisions
have a clear idea about vital points such as this, and it
is the lack of such a background that both confuses decision
making (perhaps causing decisions to be reversed at a later
date) and limits buy-in to the final conclusions. It is well
established that individuals, especially highly educated and
specialised individuals like those working in pharmaceutical
R&D, will find it difficult to go along with decisions
that they don’t understand, especially where they can
personally be profoundly affected by those decisions.
An efficient way of successfully completing this first part
of the process is to conduct one or more Delphi sessions with
the group who are doing the work to support the final decision
making. This group should ideally be made up of R&D managers
with a good understanding of the potential projects and the
risks inherent in the procedures in R&D as well as representatives
of the commercial side of the businesses. Skilled facilitation
of these meetings is vital. There is not scope here to go
into the details of group behaviour and meetings management,
but it is clear that there will be powerful opinions and individuals
in such groups whose influence must be moderated, and a good
consensus must be reached as efficiently as possible.
The objective of these meetings is to establish just what
the corporate priorities are, to answer the question “what
does the CEO want of the company’s future products?”.
All possible aims should be considered, assessed for likely
validity and grouped according to whether they are aims in
themselves or just subsidiary to achievement of other objectives.
At first all will be confusion, but gradually with good group
facilitation a set of key strategic issues will emerge. Typical
top line strategic categories would be:
Financial return (once the product is launched) -
Cost (of getting to that return)
Risk (of achieving that return) Timing (of the cost and return) Other corporate ambitions (such as desired therapeutic area focus)
Below this level there will be a number
of subsidiary factors which contribute to the achievement
of the targets set at the top level. For example, within “financial
return” there will be important elements such as:
Market size in terms of numbers of patients potentially to be treated Current unmet medical need The ability of known future competition to reduce that unmet need
Within the “risk” category
there might be:
All of these strategic factors can differ
from company to company, and skilled facilitation is required
to ensure that issues are fully aired, nothing pre-empted,
and overall numbers of factors kept down to a manageable level.
Some careful thinking is also needed to determine just how
factors interact with each other. Product pricing, for example,
may seem to be a straightforward factor within the overall
category of financial return – basically the higher
the price the better. However the question arises whether
this should be today’s price level in the specific market
concerned, which is influenced by a mix of issues such as
current unmet need, and age and quality of existing products,
or future pricing, when the planned product is likely to enter
the market. It seems clear that it is future price that should
be the important criterion. However, thought needs to be given
to whether this can be assessed as a strategic criterion itself,
or whether it is not actually already present in the “unmet
need’ criterion. It is the author’s view that
to have two elements under “financial return”,
one addressing future pricing and one addressing future unmet
need, is double counting. If there is a high level of need,
there will be the likelihood of a relatively high price. The
major drawback might be a background of historic very low
pricing, which could put contraints on the extent to which
a price which reflects the real value of the new product can
be realised. Our preferred solution therefore is in fact to
include only current price under “financial return”
because this can tell us whether such a background price constraint
exists, and to allow “unmet need” to express the
potential for a higher price profile that the product per
se represents.
This pricing example is included to demonstrate that some
careful thought is necessary in drawing up what may seem a
simple schedule.
Once these main and subsidiary categories have been determined,
they need to be weighted according to their importance within
the company. First the large scale, key factors are weighted
according to company need. (A small biotech may find that
timing of return, or of announcement of a deal, is of vital
importance, while a major corporation may feel that scale
of financial return is much more critical.) Then the subsidiary
factors are carefully weighted according to their contribution
to the key factors. These weights need to be addressed aggressively.
It is too easy to conclude that all the factors are more or
less equally important. If this is done it will not reflect
reality, and it will not permit a useful result to emerge.
Assessing project potential
Once these strategic priorities have been established, they
provide a framework for the detailed assessment of the characteristics
of each product opportunity. Qualitative and quantitative
research should be carried out into the extent to which each
potential project matches up to these requirements. As a minimum
this should involve reasonably detailed desk work into each
project area in turn. Over time expertise in such assessments
of potential can be built up. Alternatively, external advisers
can be asked to conduct the analysis, which has the benefit
of saving time and increasing objectivity. It is often very
useful to include techniques such as external interviewing
of experts in the field. Matters covered will inevitably involve
assessing the feasibility and magnitude of the project’s
development programme; the potential for aggressive competition
(taking into account both the nature of the products that
may emerge and of the companies who may be selling them);
the scale of the market and the level of clinical need; promotional,
pricing and reimbursement practice in the sector.
It should be clear from this description why the strategic
factors are dealt with first. It would otherwise not be possible
to know which elements of the various projects should be analysed
most carefully, leading to wasting of time on unnecessary
issues, and neglect of important ones.
If the projects in question are at a very early stage, there
is generally nothing to be gained by attempting full scale
financial modeling, for example to derive project NPVs, at
this point. Such models will only be based on the same set
of largely qualitative assumptions that will drive the scoring
process described, so will provide no new information and
can lend a spurious sense of accuracy to the data they generate.
It would also be very time consuming to create models across
a large number of potential projects.
The purpose of this part of the process is to provide good
qualitative data which can be used to provide reasonable scores
for the projects against the strategic targets.
Scoring and evaluating projects
To complete the picture, the projects should be scored for
their match with requirements in each strategic sub-category.
To do this, a set of questions needs to be set up which will
provide the score required. For example a scale would be established
for patient numbers which permitted the score for each project
to be read off – if there are world-wide 3 million potential
patients then the score is x, if there are 10 million the
score is y. Unmet need might have a scale relating to a composite
of the patient suffering involved and the social cost incurred
(they need not, regrettably, be the same thing, and social
cost tends to win).
Normally these scores would be allocated by those performing
the analysis of the potential demand for the products concerned.
However they would have to be presented and explained to the
team responsible for making prioritisation recommendations
and challenged by them where necessary.
The end effect of this process would be a table looking something
like the following (extract showing a few projects from a
fictitous project range):
From this data it is then possible to obtain an overall score
for each project reviewed as well as to re-examine the results
according to the strategic categories assessed, as shown in
the following charts:
In these cases, for example, projects C and E obtain their
good scores from a range of positive characteristics. They
have good returns, involve low risks and relatively low costs
(“cost” is shown in reverse in the chart –
a small circle means low cost) and even their timing is OK.
All things considered these should therefore clearly form
part of any selected portfolio.
The third best project overall, Project F, does however raise
some questions in spite of its strength in total. It gains
its high marks from its good risk and cost profile, but is
low on return and not very good on timing. It would probably
be selected, but would need to be in a balanced portfolio
that compensated in other ways for its weaknesses.
Project B is an interesting challenge, because its risk profile
is the worst of the group, it is expensive to make, its timing
is not very good, however its return is third best.
Projects A and D have few redeeming features.
By cross-analysing the results in this way, more can be learned
about the shape of the dilemma with which the company is faced.
It also challenges both the strategic weightings that were
initially chosen and the way the projects were individually
assessed against the strategic criteria. If at this stage
the results are counter-intuitive either the company has learned
something new, valuable and unexpected and made significant
steps towards a vital portfolio decision or it has learned
that there is something wrong with the way it is setting strategic
priorities and it should go back and start again. Either outcome
represents real progress.
Conclusions
The method described above is only one of several that can
be used to decide on the priorities across a portfolio of
pharmaceutical development projects. It is also one of the
simpler approaches available.
This simplicity however has shown itself to be one of its
strengths, together with the open, discussion-based approach
that involves all key management staff in setting the criteria
for the decision and then in making (or making recommendations
for) it. There is substantially more buy-in to an approach
that is as open and clearly understandable as this compared
to a relatively more sophisticated approach which uses more
complex techniques and can tend to have a “black box”
character to it – data is fed in to a procedure, and
an answer eventually emerges which may seem to bear little
relationship to the initial input, and it is not at all clear
why that is the answer and not something else.
As well as gaining the buy-in of staff involved, the procedure
has been shown by experience to have the additional merit
of exposing staff who might normally be quite far removed
from to the forces that are at work on the company to a number
of key corporate policy drivers.
Top
|
