The power law relationship in drug development

All opinions are my own and do not necessarily reflect those of Novo Nordisk.

A few weeks ago a friend and I had the great opportunity to go see Nate Silver speak at the University of Washington. He’s a funny, engaging speaker, and for someone like me who makes his living generating and analyzing data, Silver’s work in sports, politics and other fields has been inspirational.  Much of his talk covered elements of his book, The Signal and the Noise, which I read over a year ago. It was good to get a refresher. One of the elements that particularly struck me this time around, to the point that I took a picture of his slide, was the concept of the power law and its empirical relationship to so many of the phenomena we deal with in life.

Nate Silver graph small

Figure 1: Slide from Nate Silver’s talk demonstrating the power law relationship in business–how often the last 20% of accuracy (or quality or sales or…) comes from the last 80% of effort.

Because I spend way too much time thinking about the business of drug development, I started thinking of how this concept applies to our industry and specifically the problem the industry is facing with creating innovative medicines.

The Power Law relationship is one in which one of two linked elements varies as a function of the power of the other.  In The Signal and the Noise, Silver described how this simple behavior characterizes a wide variety of things, behaviors, activities… For example, the occurrence and magnitudes of floods and earthquakes, or the distribution of intensities of storms over time. One of the vexing things about a power law relationship, however, is that it does not provide a great deal of forecasting ability. Just because a magnitude 8 earthquake has struck does not mean another cannot happen within a specific number of years; only that over time, the distribution of magnitude 8 earthquakes will generally be an order of magnitude sparser than the distribution of those of magnitude seven.

Drug development empirically seems to be following a power law distribution in the relationship of effort to creation of new drugs; at this point in time, even incremental improvements in the efficacy of new drugs are taking much more time and money than improvements have in the past.  So, in drug development today we see the 80-20 rule (or maybe it’s 90-10, or worse), roughly, where an extra 10-20 % of improvement requires four times as much effort as it took to get a good therapeutic in the first place. This is a different way of framing the problem of innovation in Biopharma, which I described via the metaphor of the Adaptive Landscape a post or so ago.

This problem, which was the issue underlying the Biopharma patent cliff some years ago, and which is seeing current expression in the new round of mergers and acquisitions that seem to be on the horizon (Valeant and Pfizer, I’m looking at you), is real. While some may choose to believe conspiracy theories about the drug industry and governments somehow colluding to keep life-saving treatments from patients, the truth is that drug development is pretty tough right now. Companies don’t engage in large scale, disruptive and morale-sapping mergers unless they feel they must to keep up. Well, I assume so anyway. And it is just keeping up, just running the Red Queen’s race.

Another expression of this problem can be seen in the recent controversy over the price of Sovaldi, a truly revolutionary drug for Hepatitis C that Gilead recently brought to the market. Here is a drug that can cure an individual of Hepatitis C, in 90% of patients, removing the risk of developing liver cirrhosis or cancer and the eventual need for a transplant. By any measure this is not just an incremental improvement; it’s a true innovation. And it has a price of $84,000 per patient.

Over at the blog Drug Baron, David Grainger has a really thoughtful, almost philosophical series of posts about drug prices and how they’re set. He contrasts the concepts of “cost-plus” versus “value-based” pricing. The former is the cost of drug development with a premium for amortized cost of development. The latter may be thought of as the Platonic price of substantial, novel improvements to health and well-being. Figuring out the correct cost-plus pricing is a bit of a mess, partly because the power law relationship in drug development means drug development costs are moving targets. Value-based pricing, on the other hand, can be something a society can simply decide. However, there’s the economic reality that society may not be able to afford the price either way. There is no natural law decreeing that society will always have enough money to support every need of its citizens.

Kaiser Health News recently reported on Sovaldi and calculated that if every person in the US with Hepatitis C were treated, it would double the amount US citizens pay on prescription drugs from $300 to $600 Billion. It’s unlikely that every person will be treated immediately, and there are other Hepatitis C drugs on the cusp of release that may lead to market competition and lower prices. However, this specific drug is not an isolated product. There are many drugs moving forward in Biopharma pipelines that could well be priced similarly and be aimed at these kinds of larger markets.

The power-law relationship implies that, barring a dramatic, unexpected breakthrough in how we develop drugs, it is probable that prices will continue to rise per drug because the next generation of drugs will cost much more to develop. And companies, seeing that the drug development process will only get harder, may proactively seek to bake in future costs, not just those from the past.

So what are the implications for a power-law relationship in Biopharma and drug development? Well, for one thing, it provides economic and business support for what Pfizer has been doing in trying to purchase Astra-Zeneca. If we believe that truly disruptive innovations that reset the power-law baseline are essentially unpredictable and may only come along once in a very long while, a company may be best served by just treading water and hoping the disruption occurs before it goes bankrupt. Also, that every surviving company should actively be looking for disruption and be poised to jump into the new wave in time. Of course the history of the Innovator’s Dilemma suggests that’s not very easy for any established company to do.

Mind you, I still think what Pfizer was trying to do is pretty unpleasant from the standpoint of general innovation and employees.

A power law distribution for improvements in drug development also suggests that companies could be better about proactively managing expectations about future growth. Admittedly this is hard to do when shareholders want continual rises in share price. But it would certainly help to change the messaging. Maybe it’d be like the current trend in TV series, where it’s known from the start that a given program will have a set number of years to tell its story. Game of Thrones as a business model. Okay, maybe that’s not the best analogy.

Another interpretation might be that now that companies are at the upper edge of the curve, they should just stop trying to do drug development. Instead, they could take their resources and their sizable R&D groups and set them loose on the problem of improving human health, period, without requiring the answer to come in the form of a tidy little pill, while still trying to create a viable business model.  Trying, in essence, to go “all in” on finding that disruption that will reset the baseline. Of course, this would be the equivalent of suicide for most (all?) of these companies, since the nature of innovation is that it’s unpredictable and so may not appear despite the best efforts of all the scientists, engineers and health care professionals working for a company combined. Still, such an action could play a role akin to that of nurse logs in the forest, companies sacrificing themselves so dozens of seedlings can take root, replenishing the health innovation ecosystem.

Never going to happen, of course, and for good reasons. But it’s fun to speculate and dream.


One thought on “The power law relationship in drug development

  1. Pingback: It’s time for biopharma to embrace public health | Biotech, Baseball, Big Data, Business, Biology...

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s