By looking at the universe of active Paragraph IV cases, companies can establish timing probabilities that will show the likelihood
of a court decision at a certain time. Exactly when a court case will be resolved depends on several factors:
- The jurisdiction— Do cases in Philadelphia typically finish quicker than those in Chicago?
- The judge— Does one judge rule more quickly than others?
- How many companies are involved— Does a case involving two companies get decided quicker than a case involving five companies?
- How complicated the case is— Does it involve interpreting six patents or just one?
The end date is critically important, because if the brand company loses the patent challenge, the generics company can bring
its product to the market immediately. The data show that more generics companies are deciding to launch their products directly
after they win at trial and not wait for a court of appeals decision.
In examining Paragraph IV cases, patterns begin to emerge that help answer the question of when a certain court case will
likely end. And each individual product has particular factors that can be used to help determine probabilities of when its
particular case will end. These unique factors include:
- how many challenges it has received
- where the case or cases are pending
- the particular market dynamics
- when its patents expire
- other factors that drive portfolio decisions
These statistical techniques are reasonably reliable. Even though each case has unique factors, Paragraph IV challenges
nonetheless have many things in common, including the fact that they all litigate under the same federal law, under the same
procedural rules, and with the same types of evidence. Such commonality leads to statistical normal distribution curves once
samples are tested.
For the Brand A hypothetical example, assume that the historical data inputs, probability distribution, and weighted factors
yield an expectation that this case will be decided about 32 months after it starts, with a standard deviation of five months.
So, for Brand A, the team can figure that there is a strong likelihood that it will receive a court decision during that time
frame and it can strive to manage the product around this time frame. In addition, the other key factor for the company will
be to weigh the management of Brand A against the other brands in the portfolio.
One option is for the Brand A team to choose a hedging strategy that would reduce the number of its marketing and sales support
for a year, then have that sales support return if the company successfully defends it patents. With the timing probabilities,
it may want to reduce the sales force to half (which will slightly reduce market share) for a year starting 25 months after
the patent case begins. By doing so, it benefits by maintaining sales while reducing the risk of loss if it were to lose the
In terms of financial gain by this hedging strategy, the net present value would be $3.06 billion (if it were to win and afterwards
increase its sales force to prior levels) or $1.65 billion (if it were to lose the patent defense and face generic competition
in 2008). By using a hedging strategy, the difference between the best case (win) and worst case (lose) is $1.41 billion,
in terms of NPV.
Before, without applying the timing probability data, the spread between the best and worst outcome was much greater at $1.73
billion ($3.31 billion minus $1.58 billion). By managing around the time frame, the Brand A team can reduce the overall risk
to the brand and company by roughly $320 million. Note that only two cost items were considered in this example, and tweaking
more cost items can yield more (or less) significant differences. Also missing from this example are other financial factors
that can come into play, such as:
- portfolio issues— Will removing support from Brand A help Brand B or C?
- manufacturing— Will knowing the timing probabilities increase manufacturing efficiencies?
- general expenses— headquarters, headcount, and R&D.