Reference: BioSpace
Many pharmaceutical companies are losing revenue due to poor customer data management, according to a recent survey from Model N, which helps clients optimize revenue. Specifically, as much as 6% of pharma revenue leaks out in the form of penalties, incorrect pricing and overpaid rebates. And nearly half of the 304 executives who participated in this year’s survey indicated that their membership data management process is inefficient.
The inability to vet formularies—lists of drugs covered by health insurance—was cited as a challenge by 94% of survey respondents, and poor membership data management was a problem for 95% of respondents. These two factors are key issues impacting revenue management in the pharmaceutical industry, according to the report. Respondents listed data accuracy as the topmost issue preventing proper management of customer and membership data. Inaccurate customer data can lead to poor pricing decisions, incorrect rebate payments, revenue loss and even government penalties, according to the report.
The Roots of the Problem
Jesse Mendelsohn, senior vice president at Model N, told BioSpace that it is common for customers to register details such as names, addresses and identifiers (unique customer markers) that do not correspond across all company records, making it difficult to determine which programs they qualify for. For instance, an entity may present a purchase order with its Health Industry Number identifier, but the system only knows that customer by its Drug Enforcement Administration identifier.
Another problem is monitoring formularies for accuracy. Forty-seven percent of the survey respondents indicated that they have limited resources to vet their products on formularies, 48% said they have limited access to formularies and 49% admitted that their companies do not have the manpower required to validate formularies against the contracts they have with the long list of group purchasing organizations (GPOs)—entities that pool the buying power of multiple members (hospitals, pharmacies and clinics) to secure discounts.
Mendelsohn explained that most purchases in the industry are made through these groups and that some manufacturers do not verify whether their customers are eligible for chargebacks before reimbursing the claims. “If you are inaccurately depicting what the members of a group are, or assuming members are still active in a group when they are no longer active, you could be giving customers discounts that they don’t actually deserve,” he told BioSpace. He said it is common for companies to lose track of customer records, especially large companies that deal with many GPOs. This ultimately leads to revenue leakage.
AI Is Not a Silver Bullet
When it comes to solutions, 60% of the survey participants believed advanced analytics and AI can help them overcome their data management problem and optimize revenue. But experts told BioSpace that would be putting the cart before the horse.
“It would be nice if you could install an AI software and say ‘reduce revenue and manage my data,’ but there isn’t any technology like that right now,” said Adam Marko, director of life sciences at Panasas, which provides data solutions for the life science industry. “If your data management isn’t in place, the best AI models in the world aren’t going to help.”
Mert Zorlular, chief financial officer at the drug distribution company Er-Kim Pharmaceuticals, said that before employing automation, companies need to have “clean and usable or machine-possible data.” After this, a data lake—a centralized storage unit for all company data—can be developed to aid analysis. “The use of AI and ML [machine learning] in data cleaning and [in fixing] data errors are increasingly easy to do, and the key to accurate analysis is having a very robust data lake,” Zorlular said.
A Cost-Effective Solution to Data Quality Issues
Zorlular told BioSpace that in the life science industry, data is scattered in different places. Er-Kim Pharmaceuticals uses a data lake that collects all kinds of customer data from more than 100 sources and that this data is periodically transformed and cleaned before use to ensure data accuracy.
But this process can be very expensive, with the cost varying by company size. Zorlular said that large pharmaceutical companies with commensurately sized data sets could be looking at data lake costs as high as $25 million per year. The number is much lower for smaller companies with lower data volume, but they could have diminishing returns on investment. He advised smaller companies to invest in a team that will create a data lake for the company to help manage their data needs instead of buying expensive data analytics services. “I think that’s the key: a small but efficient team and a company culture that understands how valuable [data management] is.” Zorlular noted that the benefits of engaging in efficient data management practices generally outweigh the costs.
Mendelsohn said small companies that cannot afford to invest in expensive customer data management projects ignore the revenue leaks altogether. To solve this problem, he suggested that the pharmaceutical industry work together on a centralized data system. Manufacturers are generally selling to the same set of customers, so “a quality view of who the customers are and of which groups they are accurately a member . . . would be huge for the pharmaceutical industry.”
He noted that large pharmaceutical companies will also benefit from a centralized service that analyzes and creates clean customer data, as many of these companies are spending huge amounts of money to manage their data. Yet according to the Model N survey, only about 4 in 10 executives rely on data to make decisions, suggesting they are not getting the most from their data management practices.
Mendelsohn said that a collective view of all of the pharma industry’s customers, including their GPO memberships, “would be a huge benefit for pharmaceutical manufacturers, as they would now have the ability to more accurately see who is out there, and also what price they’re eligible for.”