Magazine Article | October 1, 2015

Does Pharma Demand Forecasting Keep You Awake At Night?

Source: Life Science Leader

By James Mullen, CEO, Patheon N.V.

Given the uniquely long and winding road pharmaceuticals must take to market, it is difficult for developers to predict the manufacturing capacity they will need when their product finally gets there. Necessity demands that developers begin thinking about production capacity around three years prior to launch, but at that time a company has no real idea if its product will be a success, or how big a success, and therefore how much capacity it will need. Will it need a million units, 10 million, or 50 million? Guess wrong one way, and the company has a white elephant and a massive loss on its investment. Guess wrong the other way, and it misses out on irreplaceable revenues.

To estimate future needs, pharma companies develop forecasts, and often quite sophisticated ones, too. But, by definition, forecasts are never 100 percent right. And it’s especially difficult to predict sales in markets that are likely to see the introduction of numerous competing products.

One way to solve that problem and relieve that uncertainty might be for a developer to outsource production to a CDMO. But those of us on this side of the fence historically have behaved as though it is our clients’ job to tell us how much product they will need. As long as we think that way, the risks of over-and under-capacity remain, threatening both revenues and profits.

Instead of forever seeking more-certain forecasts, I believe we should be talking about how to provide flexible, scalable capacity that can accommodate the uncertainty. With sufficient flexibility, the need to accurately forecast demand for a product that does not yet exist is relaxed.

There are several ways to provide flexible manufacturing capacity cost-effectively. Developers, especially larger ones, can do it themselves, and I believe CDMOs also are well positioned to do the job.

The first requirement is scale. If you have many plants and lines, it is pretty simple to accommodate unexpected demand by validating more product lines in more locations than you expect you will need. That capacity can be used for another product if that demand does not materialize. The manufacturer can distribute and mitigate the risk of over- or under-capacity for any one product.

It is also relatively easy to provide flexible capacity for biologic drug substances. For a new biologic, a typical forecast range might be 2 million to 10 million units a year. If you build capacity for 10 million units and the demand turns out to be 2 million, you will pay far more overhead than you need. If the launch is a runaway success and demand rises to 15 million units, then you forgo 5 million units a year in sales at say, 90 percent gross margin. However, with multiple single-use 2,000L reactors, you can flex capacity as you need it. Single-use technology does not require revalidation of additional reactors because the process is exactly the same in each, and the ramp-up time for an additional process is short. So capacity can be easily flexed within a facility or across more than one location.

Flexible capacity is not so easy to achieve for small molecules, as each has its technical challenges and requires certain assets in every location. One way around this (other than to invest in assets that may not be needed) is to recruit third-party manufacturers. There are many of these that make just a few products for a few clients. A large manufacturer that already has strong quality and technology transfer capabilities can assimilate such companies into its network and impose its standards and ensure scalable capacity, quality, and delivery.

The pharmaceutical industry is in a period of rapid change. Scientific and technological breakthroughs, changing regulatory requirements, payer pressure, patient demands, and rising costs are all leading to more challenges for developers. To focus on these challenges, it is important we eliminate whatever other risks we can. I think that demand risk is eminently manageable at a reasonable cost. And I think we’ll all fare better if we think of it as a risk we can manage by flexing capacity, rather than one we can eliminate with a better crystal ball.