Richard Gray, commercial director at Syrris, highlights the paradox that despite the constant acceleration of the drug development process, the list of new drug registrations dwindles each year
Many tonnes of paper have been covered with articles which try to find ways to reverse this decline in drug discovery productivity.
Does this mean that the decline is real and fundamental? This is possible, given the duration of the trend, and the hundreds of billions spent failing to reverse it, so we need to look at how the industry will try to adapt to face this challenge.
One approach would be to stop the focus on 'accelerating', and to start focusing on 'arriving sooner'.
The core issues for the pharmaceutical industry are:.
* The easy problems have been solved.
* The useful economic life of any solution is being eroded by competitors.
* The task is being made harder by ever increasing regulatory pressures In today's information-rich world, we need to take the time to absorb the information we are receiving, think about the implications, and then act decisively upon them by modifying strategy and direction.
In the world of pharmaceutical process development, the scientist needs to gather teams, tools and processes that value and foster agility rather than raw speed; that allow rapid collation, presentation and processing of data; and that focus on learning the most from each set of experiments before moving on to the next.
In the words of one of the scientists who worked on the development of Viagra: "...it will be important to integrate new scientific advances into an environment that builds on traditional skills, (and) fosters multidisciplinary interactions between teams and individuals." (Science, art and drug discovery: a personal perspective; Simon F Campbell, 2000).
Drug discovery is essentially a field of huge opportunity and almost boundless possibilities; there are perhaps 1x10^40 possible drug-like molecules, and tens of thousands of disease targets.
The accepted view of the landscape in which discovery is played out is portrayed as boundless, lush, and rich with opportunities.
In these environments, there is plenty for all, and a high degree of discovery diversity.
However, if the pharmaceutical landscape was not so, then it would be a better strategy to focus on competitive edge, methodology, and processes.
Thus, only time will tell if scientific advances can open up the full potential of the genome and the proteome to myriad therapies.
Until that happens, a wise strategy may be to assume resources, ideas, and solutions are much more scarce.
Reaching the goals The goal of pharmaceutical process development is straightforward: "The aim of pharmaceutical development is to design a quality product and its manufacturing process to consistently deliver the intended performance of the product.
"The information and knowledge gained from pharmaceutical development studies and manufacturing experience provide scientific understanding to support the establishment of the design space, specifications, and manufacturing controls" (European Medicines Agency, Note for guidance on Pharmaceutical Development, 2004).
Many of the terms used in this definition lead to clear targets - eg, "design a quality product and process to consistently deliver intended performance".
However, the definition is also relatively open in deciding how to approach many key areas.
Exactly what "information and knowledge" is required from which "development studies", and how is "scientific understanding" irrefutably demonstrated? Most importantly, how can this overall goal be reached more quickly and efficiently? The key here is to match the tools and processes used to the targets that lie behind the statement above.
The EMEA and the FDA are increasingly moving from 'static' methodologies (those where the end-point is defined and validated) to more 'dynamic' validation methods (where the process is fundamentally stable, measurable, and controllable at all times).
Recent examples of this trend include:.
* Continuous process verification.
* Quality by Design.
* Process Analytical Technology - a system for designing, analysing, and controlling manufacturing through timely measurements (ie, during processing) of critical quality and performance attributes of raw and in-process materials and processes with the goal of ensuring final product quality (FDA Office of Pharmaceutical Science, 2005) Hence the scientist must adapt working methods to address these goals.
The changes required are both fundamental, in terms of the level of change required, and also in the impact on all stages of process development.
This is because dynamic validation is much more fundamental and robust, and it therefore will expose weaknesses very clearly.
Hence the "searchlight" effect of these techniques needs to be considered very early in the development process.
Winning the relay race.
The relay race required to deliver efficient process development needs a set of excellent resources.
The characteristics of the successful scientist and of successful tools for the scientist to use are typically similar.
Both must be interactive, integrated, agile, and flexible.
Let us focus more specifically on the products and tools used.
In addition to the characteristics above, successful products must also have low entry barriers to use, be suitable for early adoption, and as far as possible, must pull traditionally late processes into the earliest possible steps in the development program.
This lets the process 'fail early' - something of a mantra in current thinking, as quite rightly it emphasises the huge waste that can be avoided by finding out bad news early on.
It's important not to forget the hidden second part to that philosophy, which is not so often mentioned; fail often.
Some authors use the term 'churn' to describe the creative process of forming ideas, testing them, and creating new, better models when the early ones fail (Guy Kawasaki, Rules for Revolutionaries, 2000).
In successful process development, the goal is to continually churn and try to validate models to describe the process and the product.
The scientist must generate models and adapt them when they fail, and with each error comes an increased chance of success in the next trial.
So, according to this model, optimal chance of success is a function of frequency of trial, and not the speed of the trial - going round in a circle one more time may actually get you there more quickly.
Failing often means testing often, and testing early.
Both imply a move towards more automated testing processes, and ideally a need for clear rules and tests to determine suitability.
To give some tangible examples of recent advances in this area:.
* Polymorph screening instruments are increasingly in use at early stages in the development process, using high throughput parallel systems to identify and characterise the crystal structure of the target molecule.
* Calorimetry is being moved out of the specialist lab.
New techniques such as constant flux calorimetry (developed by Ashe Morris) give the ease of use and instantaneous data required to give every chemist deeper insight into the thermal processes occurring during product synthesis.
* Flow synthesis and purification technologies, eg from Syrris and other vendors, are being used to offer rapid process optimisation with small amounts of material, coupled with the potential for seamless scalability to kilo production and beyond.
*Green chemistry techniques are being promoted to offer environmentally friendly processes that also offer sound science, and reduced energy consumption.
The techniques include resurgence in the use of solid phase reagents and catalysts (as promoted by Prof Steve Ley and others (eg Chem Commun, 2006, 2566 - 2568)) to give fewer more efficient process steps, which offer close to quantitative yields with little need for inefficient and solvent-thirsty separations.
In conclusion, it is clear that to work efficiently and optimally, the process development scientist needs to trust the time-proven scientific method: form a hypothesis, test it, and when it fails, improve it and try again.
The method is no less valuable in the current environment, and it is likely to prove more efficient than more 'linear' techniques at delivering rapid convergence towards a successful product development goal.
Coupled with new adaptive and agile tools and methodologies, there is excellent scope for the approach to deliver robust processes for a new generation of drugs.
The true test of these techniques will come if it becomes more evident that truly new drugs are indeed rare, elusive, and hard won.
Ever-reducing drug registrations seems to be one clear indicator that this may well be the case.
The challenge to the industry is to improve its science and methodologies to increase the number of project failures early on, rather than wasting time/cost on developing a drug that fails late on in the process.
It is far quicker to reach today's target than to continually race towards a receding goal.
The model suggested is therefore to use a 'fail early and often' approach in order to weed out poor drugs, but not to give up too early solely because of processing difficulties.
Process chemists should work to use the expanding array of process development and analytical tools to seek solutions to manufacturing or processing problems that would in earlier times have led to rejection of a promising drug candidate.
By combining a 'fail early' testing approach with a 'solve early' process development process, a larger number of promising drugs may reach the clinical trial stage.
A robust process development methodology improves the chance of the drug reaching the market; better tools and better methodologies help lead to better drugs.
The earlier the intervention, the more likely the outcome will be successful.
