7.10.2008

Commentary: Pharma's future is virtual, say Kate Moss and Dr Steve Arlington of PricewaterhouseCoopers

As pharma switches from developing palliatives to cures, its productivity in the lab is
plummeting. The number of innovative new medicines reaching the market is falling,
while the time and money required to research and develop them continue to soar, with
major strategic implications for the industry as a whole, says PricewaterhouseCoopers.
In its latest Pharma 2020: Virtual R&D report on the future of the sector,
PricewaterhouseCoopers argues that incremental changes are no longer enough.
Companies will need to decide what they want to focus on - whether, for example, to
continue playing in a mass market or concentrate on developing speciality treatments.
And whatever route they choose, they will have to make fundamental changes in the way
they work.

virtual humans

It is now widely recognised that, if pharma is to produce therapies which deliver the
value patients, payers and providers seek, it must acquire a much better grasp of how
human beings operate at the molecular level and the functional changes associated with,
or arising from, disease.
This knowledge can then be used to create predictive computer models for simulating the
physiological effects of interacting with specific targets, identifying which targets are
involved in a given disease and determining what sort of intervention is required.
This is a huge task. Numerous organisations are already building discrete models of
different organs and cells, but these models must ultimately be integrated into a single
validated model in order to predict the effects of modulating a biological target on the
whole system. That model must also be capable of reflecting common genetic and
phenotypic variations, so the computing power required to support it will be enormous.
Nevertheless, various efforts to create a digital representation of the human body are
already underway. The Step Consortium is developing a technological structure for
investigating the human body as a single system. The Living Human Project is building a
model of the musculoskeletal apparatus; and the Physiome Project is creating a
computational framework for understanding the integrative function of cells, organs and
organisms.
It is unlikely that "virtual man" will be available by 2020, but virtual cells, organs and
animals will certainly play a much bigger role in pharmaceutical R&D. Several other
technologies will also make a major contribution to the process. Semantic technologies
will, for example, enable scientists to connect disparate data sets and query the data
using "natural language". This will make it much easier to identify previously
unobservable links between a particular disease or molecule and the pathways it affects.
Similarly, advances in nanotechnology and genetic engineering will enable the industry
to develop totally new forms of treatment which improve compliance and hence
outcomes.
Once virtual man has been built, it will also be possible to screen candidate medicines in
a digital representation of the human body which can be adjusted to reflect common
genetic variations and disease traits. This will show whether a molecule interacts with
any unwanted targets and produces any side effects, and in what circumstances it does
so. Predictive analysis will then enable researchers to assess how the molecule is likely
to be absorbed, distributed, metabolised and excreted; what long-term side effects it
might have; what free plasma concentration is needed to provide the optimal balance
between efficacy and safety; and what formulation and dosing levels might work best.
one phase for clinical development
Of course, even the most robustly modelled molecules will still have to be tested in real
people. However, the development process will also change dramatically. Clinical trials
are currently cumbersome, time-consuming, require the participation of numerous
patients and often yield unclear results. Much of this inefficiency stems from the fact
that patient data resides in different formats in multiple databases, which cannot
communicate. Without easy access to information that is recorded consistently, the
process of designing trials and monitoring patients remains much as it was 20 years ago.
But common data standards are now being developed and many countries will introduce
electronic health records within the next decade. Semantic technologies will also enable
pharma to link trial data with epidemiological and early research data, and identify any
significant patterns, while pervasive monitoring will let it track patients on a real-time
basis wherever they are. The new European Centre for Connected Health is one of
several organisations already testing various remote monitoring systems, and new
technologies will facilitate the creation of embedded devices, such as electronic circuits
which can be "printed" on the skin to track a patient's blood capillaries and nerve
endings.
These advances will ultimately render the current model of development, with its four
distinct phases, obsolete. The process will start with the administration of a treatment to
a single patient, who has been screened to ensure that he or she has the relevant medical
profile. Once there is evidence that the treatment does not cause any immediate adverse
reaction, it will be sequentially administered to between 20-100 more patients, all of
whom have also been screened. The data they generate will then be compared with data
from the modelling that preceded the study and subjected to techniques like Bayesian
analysis. These findings will determine how the study is adapted, but the study itself will
be conducted in a single, continuous phase.
The needs of patients, payers, providers and the regulators will also play a much more
prominent part in shaping how trials are designed. Many companies submit products
which, although they secure regulatory approval, have limited therapeutic value and fail
to generate a commercial return. But, by 2020, pharma will collaborate with payers
when they develop their trial protocols to assess the value of new treatments.

automated approvals

Pervasive monitoring and electronic health records will transform the approval process,
too. At present, a regulator reviews the data on a molecule at the end of development,
when the sponsoring company submits the supporting dossier. By 2020, this all-ornothing
system will be replaced by a cumulative approach.
The sponsoring company will collaborate with the regulator to establish the evidence
that is required and then submit it electronically, in line with a predetermined schedule.
Once sufficient data have been collected to show that a medicine is safe, efficacious and
cost-effective in the initial study population, the regulator will issue a "live licence"
allowing the company to market the treatment on a restricted basis, subject to the
collection of further data on how it performs. With each incremental increase in evidence
of safety, efficacy and value, the regulator will extend the licence to cover more
patients, different indications or different formulations.
A precedent for contingent licensing has already been established with orphan drugs, and
by 2020 sophisticated monitoring devices will produce a day-to-day environment that
equates with the controlled environment in which trials are conducted today. This
process also has several practical benefits. It will enable companies to begin recouping
their costs more quickly, thereby helping to fund further testing of new products in larger
populations and the development of new therapies for different subtypes of the same
disease. It will also enable the regulators to manage their resources more effectively,
since they will be able to forecast their workload much more accurately.
conclusions
The changes we have outlined above represent a seismic shift in the way pharma
operates. The industry leaders will have to begin by deciding what sort of companies
they want to be and aligning their business models accordingly, to ensure that they have
the right corporate culture and skills to realise that vision.
They will also have to collaborate much more extensively - with each other, with
technology providers and with everyone in the healthcare value chain. But those that
succeed should get a fair return for their efforts. If society stifles pharmaceutical
investment in R&D by refusing to pay for real innovation, who will produce the new
medicines we need?

Kate Moss is a director in the firm's Pharmaceutical and Life Sciences Industry Practice.
Dr Steve Arlington is the Global Pharmaceutical and Life Sciences Industry Advisory
Leader of PricewaterhouseCoopers. SCRIP - World Pharmaceutical News -
www.scrippharma.com FILED 09 July 2008 COPYRIGHT Informa UK Ltd 2008

2 comments:

Anonymous said...

Everytime I read this article I chuckle. It is a non-scientists view of the power of modeling. We don't understand how many current drugs work so the idea that we could develop disease models to test new drugs is silly.

D

ustaginnus@hotmail.com said...

yes, there is so much truth to it... economists (at least the Austrian school would call it "prtenetion of knowledege (Ansmassung von Wissen) or more simply: we can not compute the world...