Drug Baron

December 10, 2013
by admin

Monte Carlo models of drug R&D focus attention on cutting costs – Part 1

There is a theme behind many of DrugBaron’s musings over the last three years: pharma R&D is just too expensive to make economic sense.  Given high failure rates throughout the process, including in particular a significant rate of late stage failures when the capital at risk is very high, either attrition must fall or costs must come down.

Almost everyone in the industry recognizes this equation.  But for most, particularly those who are guardians of large (and expensive) R&D infrastructure, it has been more palatable to talk of improving success rates than decreasing costs.

What cost cutting there has been has been quantitatively and qualitatively wrong.  Pruning a few percentage points off R&D budgets that have tripled in just a little over a decade has no discernible impact on the overall economics of drug discovery and development.  And cutting costs by reducing the number of projects, rather than reducing the cost per project, is not only ineffective but counter-productive as DrugBaron has already noted, on more than one occasion.

But there is a fundamental tension in the equation: success rates are assumed to be heavily tied to expenditure.  If you spend less per project, attrition rates will go up (assuming at least a proportion of the money is being wisely spent) and you will not improve the overall economics.  You might even make it worse.

So what makes DrugBaron so confident that dramatically cutting the cost per project makes sense? That even if decision quality declines slightly, it will be offset by a greater gain in productivity?

The “evidence” comes from sophisticated computer simulations of early stage drug development that underpin the ‘asset-centric’ investment model at Index Ventures.  Models that have remained unpublished – until now.

Continue Reading →

December 9, 2013
by admin

Monte Carlo models of drug R&D focus attention on cutting costs – Part 2 (the caveat!)

“So DrugBaron tells me that drug discovery and early stage development is essentially random, and so all I have do is cut costs to the bone.  Then if I have enough shots on goal, even if I don’t really know what Im doing, then I will hit big after a while.  In other words, drug discovery is all about having the right model rather than the right people. Great!”

Dead wrong.  Doing drug discovery with the costs pared to the bone is definitely a job for the experts

It is actually much harder than doing drug discovery and early development the conventional, high cost, way.  Cutting costs means doing less experiments – which means basing your decisions on less data.  Rigorously selecting, then, which are the pieces of data that yield the most actionable data per dollar spent is a sophisticated ask.  Its certainly not formulaic – it differs for every different drug discovery or development program.  Only rockstar analysts with a comprehensive understanding not just of each part of the discovery and development process, but with an insight into the overall pathways development might take, has any chance of succeeding in the highly resource-constrained environment.

The protection against making the wrong selection is just to do more experiments (and spend more money).  If you take all possible development paths in parallel, you will be completely protected against making the wrong selections (since you didn’t make any, you just dialled up the “by the book” development plan) – but at the cost of generating many bits of data that were neither essential or actionable, at least at the stage in the development process when you acquired them.  More capital at risk early means lower R&D productivity.

And if designing minimum-cost discovery and development plans is a specialist job, its not the only requirement for sophisticated insight to make the model work.  Inherent in the mathematical models are two factors that require skill: firstly, the model assumes that 50% of the doomed candidate projects are eliminated before starting (mimicking the selection of projects to initiate) – something that requires experience and in-depth scientific knowledge to deliver in practice.  Second, and even more important, there is a cut-off in terms of decision quality BELOW which cost is no longer the dominant factor.  That quality threshold is somewhere around getting two decisions out of every three spot-on.  That may be the domain in which the best drug developers are working, but it is not the kind of decision quality you can expect for ordinary mortals.

Lastly, the mathematical models assume PERFECT execution.  Fallibility is built into the kill/continue decision-making process (through the introduction of false-positive and false-negative rates) BUT no failure is permitted in the action part of each action-decision pair.  It is absolutely critical to success that none of the “working” projects are destroyed by simply choosing to do the wrong experiment, or doing it so badly that it yields the wrong answer.

Poor execution will kill returns even quicker than higher costs

Rockstar drug developers, then, remain the sine qua non in this brave new world of low cost development, just as in conventional R&D infrastructures.  In fact, it is even more demanding.  Concluding that early discovery and development has a large random component, perhaps counter-intuitively, RAISES the value of the genius players even higher.  Delivering success through asset-centric investing at Index Ventures then is every bit as much about the quality of our Index Drug Developers as it is about having the right model.

Worse, if other people followed the model (drastically cutting costs) without having the talented people – and not just any talent, but the talent to make the right selections of what to do when resources become hyper-constrained – then they will perform LESS well for adopting the model than with their conventional R&D infrastructure.

Quite simply, how far costs can be cut within a given organization (following the DrugBaron principle of cutting cost per project, rather than simply reducing the number of projects to save cash) is determined by the quality of the drug developers.

DrugBaron’s simple conclusion, that the ultimate focus should be on cutting costs, needs a caveat: it should say, provided you have access to talented drug developers, experienced at working this model, then (and only then) will adopting an unrelenting focus on cutting costs per project deliver significant productivity gains.  After all, there is no substitute for genius, no free lunch for choosing the right investing model without also choosing the right assets and teams to invest in.  The “kill the losers” strategy is not a panacea for inferior quality, but a turbocharger for those possessed of the necessary talent already.

October 24, 2013
by admin

Evidence that “pick the winners” is precisely the wrong strategy

The asset-centric platform at Index Ventures is built around stringency.  The entire concept arose from the realization that drug discovery and development is a complex process, in a low-validity environment (meaning there is often too little data available to make decisions), where the amount of capital required is steeply graded from little to vast sums.

These parameters together imply a requirement for a filter that is more tolerant of false negatives (a decision to kill a project that would, if continued, succeed) than of false positives (a decision to continue a project that will fail later).   We call that stringency.

But success is not just about ratcheting up stringency – its about when to apply it as well

It turns out that the “gradient” of filter stringency with time is the major difference between the “pick the winners” strategy, that has some vocal supporters in the pharma industry, and the alternative that DrugBaron calls “kill the losers”.

Moreover, stringency is easier to apply in theory than in practice.  As DrugBaron has noted previously, cultural factors bias the kill/continue decision making process in favour of continuing – which automatically lowers the filter stringency.  And these cultural factors can be very strong indeed.

Our entire asset-centric platform has therefore been constructed to build in stringency, and to counteract the cultural bias towards continuing.  This is achieved in at least three ways: by stripping away the complexity due to having more than one asset in a company, everyone involved can see more clearly the risks and benefits of continuing with a single asset.  And at the same time, by providing an environment where entrepreneurs are protected from the consequences of the kill – so long as they have done “the right things” with an asset. And perhaps most importantly, by ensuring each entrepreneur has a single shot at a time to achieve a significant upside, they are incentivized to ask repeatedly if their current project really is the best use of that singular opportunity.  Equally importantly, its built to grade that stringency from low to high as the capital at risk increases.

But, as DrugBaron commented only recently, the benefits of such a model remain substantially theoretical – just like the arguments against “pick the winners” as a strategy.  These strategies are relatively new, so data is difficult to come by.  The asset-centric concept has only been mature from a handful of years, and while the first exits are certainly encouraging, the jury remains out as to whether asset-centric investing really can deliver a step-change in return on capital deployed. The “pick the winners” strategy in pharma is even more recent.

Against this background, it was certainly exciting to see the conclusions of a recent analysis by Boston Consulting Group (BCG) looking at factors that predicted success in pharma R&D.  It provides actual evidence for the principles of asset-centric investing, and against “pick the winners” – the key, as we suspected all along, is not only high stringency filters, but calibrating stringency in line with capital at risk.

Continue Reading →