Jeffrey Borman, Integrichain, Inc.

From 2011 to 2015, patent exclusivity will expire for drugs totaling more than $140 billion in sales. This staggering number has drawn much of the industry’s attention to the impact on top-line sales. Yet the risk of unexpected product returns — the inventory that will come back from the channel after loss of exclusivity (LoE) – has quietly become a focus for CFOs faced with blockbuster patent expirations. 

Over the past decade, CFOs were regularly surprised by the volume and duration of post-LoE returns on smaller brands. Now, with billion-dollar blockbusters facing generic competition, the problem is far more acute. If the returns reserve is off by just one week of inventory, the impact on the balance sheet can be tens of millions of dollars. In response to this risk, leading Finance, Forecasting and Trade departments are collaborating on new approaches to returns forecasting and returns optimization. 

	Most brand manufacturers maintain product returns policies whereby they will credit a pharmacy, hospital, or practice a percentage of the wholesale acquisition cost (WAC) of product that is at or near its expiration date. Typically, returns rates are quite low — on average blockbuster brands will have return rates between 1% and 2% of annual sales (Table 1). When a brand loses patent exclusivity its returns rate will rise slightly, then spike significantly for a period of time between 18 to 48 months following LoE. 

These returns occur because wholesalers, pharmacies, and other points of care still have considerable inventory of the brand when the first generics are launched. As the brand’s share of prescriptions falls, it’s difficult for the channel to use the inventory it still has. Moreover, most manufacturers’ returns policies protect the channel. Even if points of care wind up with too much inventory after LoE, they’ll still receive a credit for at least 90% of the price they originally paid for the product. 

Returns continue at their increased rates for up to four years after LoE. A typical brand has between 2.5 and 3 years of remaining dating when it reaches the channel, and can be returned for credit up to one year after its expiration. This means brands sold into inventory containing 3 years dating on the last day before LoE could be returned as much as four years later. 

The brand’s strategy in the run-up to LoE can make the returns challenge more acute. Many brands take multiple price increases in the quarters preceding LoE. This causes the channel’s inventory to appreciate. Since returns policies are usually based on the WAC of the product when it is returned (not when it is sold), the channel can even profit despite the effects of the post-LoE returns if the price increases are sufficiently steep. 

Similarly, brand strategy after LoE can cause increased returns rates. Since promotional spend is largely discontinued after LoE, the remaining sales of the brand (5% or more of the original market) are a very profitable and attractive annuity. There seems little reason not to sell into whatever demand remains, and very few commercial teams have systems in place to closely examine the source and viability of this demand. If actual demand proves lower than the channel’s post-LoE orders, channel inventory levels can actually increase in unit terms and take even longer to come back to the manufacturer. 

	Most post-LoE returns reserves are based, in part or in whole, on estimates of the amount of inventory in the channel at the time of generic launch. Historically, these estimates have been highly assumption-driven, since manufacturers have incomplete visibility to channel inventory. 

Wholesalers report their distribution center inventories to manufacturers in EDI 852 Product Activity reports. But beyond the wholesaler, inventory visibility is a challenge. Trade teams are occasionally able to get chains to provide guidance or ad hoc reports of their inventory levels. But very few chains consistently and accurately report their store inventories to manufacturers. Institutions, independents, mail pharmacies, long term care, and third party morgues are complete blind spots. 

Uncertainty about downstream inventory has a material impact on returns reserve accuracy post-LoE. While fee-for-service agreements have reduced the volatility of wholesaler inventories, brand manufacturers do not have comparable agreements to govern inventory levels in the downstream market. Retail chain inventories can be quite variable, driven either by the chain’s own balance sheet considerations or price appreciation. While much of the cyclical variability is contained to chain distribution centers, in-store inventories can also vary greatly from brand to brand. 

As Table 2 shows, retail channel inventories can vary widely by brand and by quarter. Even relatively fast moving brands ($500 MM+ in revenue) can see retail channel inventories vary from as few as 11 days of supply to as many as 40 days of supply.

	Table 3 shows how retail inventory uncertainty can impact a post-LoE reserve. The model arbitrarily assumes that 75% of retail inventory on the day of LoE will be returned for credit, at 90% of its WAC cost. The table displays how the reserve would have to vary just to cover the retail channel portion of total pipeline inventory (with four different brand revenue levels). 

As Table 3 shows, a brand with $1 billion in US sales might have to reserve anywhere from $20 million to $75 million just to cover retail channel inventory. 

	Brand Pharma CFOs are quickly realizing the importance of retail inventory visibility. Over the past two years, IntegriChain has worked with six of the top 15 US pharma manufacturers to improve Finance’s ability to monitor retail inventories, forecast post-LoE returns reserves, and justify those reserves to internal and external stakeholders.

	These six manufacturers are improving their returns forecasts by developing models that use the following inputs. 

 Wholesaler inventories can be as much as 50% of the total inventory in the channel, and EDI 852 provides a highly accurate and timely measurement of the inventory on hand in wholesale distribution centers, as well as unsalable inventory in wholesaler morgues that has not yet been returned. 

 867 data reports product sales and transfers that wholesalers submit to most manufacturers in support of fee for service/distribution agreements. 867 data is the wholesaler’s report of the specific locations to which they shipped product. IntegriChain provides services that help manufacturers to leverage 867 data to measure purchases at each point of care in the downstream channel. 

Returns forecasts use 867 data for two purposes. First, the 867 data identifies the number of pharmacies purchasing a brand facing LoE. Barring diversion (uncommon in the US), those pharmacies that do not purchase do not have inventory. Second, the 867 data identifies which chains are warehousing the brand in their own DCs. If a chain is warehousing the brand, it will typically carry at least one to two weeks of inventory in its DCs in addition to inventory in its stores. 867 data trends can help clarify where in that range the DCs’ inventories stand at time of LoE, as well as instances in which the chain’s DCs have ramped inventories due to pre-LoE price speculation. 

 The next step is to survey inventory levels at those pharmacies the 867 data determined to be purchasing the LoE brand. In the past, pharmacy surveys have been based on small samples (200 pharmacy calls out of 56,400 retail pharmacies and tens of thousands of institution pharmacies). Those older surveys were also random — a pharmacy was called without regard to whether it actually purchased the product in the first place. 

Successful LoE surveys are far larger and more targeted. While the exact survey size can be tailored to the brand’s turnover, IntegriChain has found 500 completed calls to be a minimum. Some blockbuster brands may require as many as 2,000 completed calls. But most importantly, the calls must be targeted. The above-mentioned LoE forecasts utilized a methodology of over-sampling high volume purchasers. Out of a total of 1,000 calls, 250 were made to quartile four pharmacies, 250 to quartile three, 250 to quartile two, and 250 to quartile one. If a pharmacy had not purchased in 180 days, it was not called at all. This enabled the forecasters to gross up the result set with a far lower margin of error in the largest purchase volume pharmacies — the pharmacies most likely to hold multiple pack units of inventory on the shelf. Validation exercises have consistently found this methodology to carry a +/-2 day on-hand margin of error at a national level. 

 Analog brand returns data provides a critical additional input to post-LoE returns forecasts. In the case of LoE, the analogs would be a combination of third party (EDI 180) and wholesaler returns (EDI 867) data for brands that have already lost exclusivity and seen generics enter the market. The analog data would encompass up to five years of returns data following each analog brand’s first encounter with generic competition. 

While inventory data provides a sound means for establishing the “worst case” (all channel inventory at LoE comes back), the reality is that some portion of the brand’s channel inventory at LoE will be used by the channel to satisfy lingering demand, particularly in Paragraph Four scenarios. 

If analogs are properly selected, a forecasting team can model the percentage of channel inventory that will be utilized against lingering brand demand. The forecasting team can also see the likely timing of post-LoE returns, which can vary greatly.

	Selecting the right analogs is critical. Manufacturers will almost always require analogs from outside their own portfolio, since a good analog set should be matched to the LoE brand based on: 

	* Average product dating in the pharmacy

	Having reviewed their returns forecasts, manufacturers are taking proactive actions to optimize post-LoE returns. Returns impact the organization’s bottom line, even when the forecast and reserves are accurate. Manufacturers typically credit returns at less than 100% of WAC. Among other things, the delta between the sale price and the credit could cover the manufacturing and distribution costs associated with the original sale. But the return credit will often be calculated based on the present day’s WAC (at the time of the return). The present-day WAC is likely higher than the price at the time the product was sold. As a result, the manufacturer will often pay out as much as or more than the sale price originally received for the product. 

In normal situations, with returns running at 1% of sales, many manufacturers accept this risk as a cost of doing business with the channel. But following LoE, returns rates run exceptionally high. Manufacturer price increases can also be quite aggressive. Consider the following scenario:

	* Brand X with $500 million in revenue goes through LoE

	* Brand X has 45 days of supply in the channel (wholesale + in transit + retail) at LoE, totaling ~$60 million in inventory valued based on the price as of the time it was sold

	* Returns policy pays 90% of the current WAC price

	* 25 out of the 45 days of supply are ultimately returned post-LoE, with 5 days returned one year after LoE, 5 days returned 2 years after LoE, and 15 days returned 3 years after LoE

	* The manufacturer increases price 10% per year after LoE.

	In this situation, the 5 days of inventory returned 1 year after LoE are credited for roughly the amount at which the manufacturer originally sold the product. Brand X therefore loses the original manufacturing cost, distribution cost, and fee- for-service cost — hypothetically 12% of WAC. The next 5 days lose 21% of WAC – distribution costs plus the 9% premium to the original sale. The next 15 days lose 32% of WAC –costs plus three years of price increases. Added up for Brand X, a $500-million brand, the scenario shows roughly an $8.7 million hit to profitability. 

Simply put, returns are not free. Manufacturers can and are taking steps to optimize the total volume of post-LoE returns. Some volume of returns is inevitable. Go too far in tightening channel inventory and returns policies pre-LoE, and prescriptions will be lost to stock outs. Innovative Trade teams are taking the lead in leveraging channel data, policy change, and channel relationships to improve their organizations’ exposure. Their successes include:

	1. Identifying chains with higher inventory levels, and partnering with those chains to ensure the inventory is run down pre-LoE and/or permitted as an early return immediately following LoE. This has to be done with careful attention to the brand’s forecast of its Rx retention post-LoE, as it may be best to leave the channel thoroughly stocked (for example, in Paragraph Four situations where generic inventory availability is uncertain). 

2. Identifying opportunities for chains and other major customers to return inventory to the wholesaler’s resalable inventory after LoE, instead of returning to the third party morgue by default. If the chain’s inventory goes back to the wholesaler’s inventory, the wholesaler has an opportunity to resell that product to legitimate demand. 

3. Manage down channel inventory pre-LoE. Innovators are using 867 data and retail inventory analytics to carefully manage wholesaler orders in the weeks before LoE. The goal is to minimize wholesale inventory and downstream inventory speculation while avoiding unacceptable increases in stock outs. 

4. Tightening returns policies to avoid crediting “bundled” returns. Desirable for many reasons, this step also ensures transparency following LoE, when Finance and Trade will need to be able to rule out pockets of “surprise” inventory. PC

Joshua I. Halpern (left) is Vice President of Marketing & Business Development, and Jeffrey Borman, Vice President of Advanced Analytics, at Integrichain, Inc. (Princeton, NJ; 

Articles

Overestimating product reserves as loss of exclusivity (LoE) approaches can be a $75-million mistake