A U.S. pharmaceutical industry veteran who was able to see major data integrity failures at a growing Indian drug maker only because of the high-level job he took there is advising global manufacturers on how to peer into their far-flung networks to judge their integrity remotely, much as credit card companies do to protect their customers.

Summary of Key Points

Quality units can be too detail-oriented to connect the dots.
Wave of patent expirations provided the temptation.
Knowledge management system revealed fraud internally.
Unsealed complaint revealed enormity of allegations.
FDA learned how to audit for data integrity.
The limits of consent decrees in the absence of strong corporate governance.
Why global outsourcing will decline over the next decade.
How Thakur’s firm, Medassurance, is using quality metrics to help global manufacturers monitor their networks.

If Ranbaxy Laboratories Ltd. had hired Dinesh Thakur away from Bristol-Myers Squibb to work in its quality organization, he never would have seen the widespread fraud that he found there in 2004.

“If my role was an analytical chemist, I can tell you … the systems and processes are designed in such a way that if you’re at a very focused kind of situation, it’s very hard for you to connect the dots,” he told the International Society of Pharmaceutical Engineering’s annual meeting Oct. 12 in Las Vegas.

But instead Ranbaxy brought Thakur to its Research and Development Center in Gurgaon, India, to manage its portfolio of generic drugs.

Consequently, he said, “I was unfortunately responsible for making sure that I looked at the entire supply chain all the way to market. That gave me insight into what was happening in the Indian industry.”

The stage was set for fraud

The blockbuster era of the U.S. drug market was drawing to a close. Ahead were the greatest opportunities for generic drugs since Congress had passed the Hatch Waxman Act in 1987 – and the greatest potential for fraud since the generic drug scandal that followed on the heels of that legislation.

What Thakur discovered during his 18 months at the helm of the knowledge management systems he built for Ranbaxy has informed his subsequent career as an industry consultant.

Ten years later, he took the podium as keynoter of the ISPE annual meeting to explain how he is helping pharmaceutical manufacturers learn to monitor their outsourced global networks for the type of data integrity failures he saw at Ranbaxy.

As the complaint he filed against Ranbaxy explained, Thakur established a program management office for Ranbaxy, where he compiled generic drug formulation, bioequivalence, stability and other data to coordinate the filing of abbreviated new drug applications, or ANDAs, at pivotal marketing opportunities.

He projected revenues for each drug in the company’s portfolio for four global regions, including the U.S., and established electronic systems to file ANDAs and manage clinical data.

What Thakur saw at Ranbaxy

Then he found out that the company was fabricating data on drugs it was selling in Africa. “That was the beginning,” he told the conference.

“At the instruction of my manager, I took it upon myself to investigate, to find out what was wrong with it, and then brought it up to the management of the company, who essentially decided not to do anything about it,” he explained.

The complaint that the Justice Department unsealed in May 2013 provided more details about Thakur’s investigation and his findings. It said that in reviewing the company’s entire drug portfolio, he “discovered that there was little or no underlying data or, to the extent that the data existed at all, defendants had fabricated it.”

The list of falsifications the complaint alleged went on and on. Stability studies were fabricated or used different formulations than proposed. Bioequivalence studies used innovator drugs that the company ground up, encapsulated and misrepresented as its own.

What FDA found with his help

After leaving Ranbaxy in 2005, Thakur contacted FDA’s center for drugs and shared what he knew with Edwin Rivera-Martinez, who was working in the center’s compliance office.

With Thakur’s help, the agency was able to discern irregularities in stability sample management at one of Ranbaxy’s facilities. The narrow scope of the warning letter that followed was a testament to the difficulty of data integrity investigations Also see "Time Waits for No Lab, Ranbaxy Learns" - Pink Sheet, 1 March, 2009..

Eventually, however, the agency learned enough to invoke its application integrity policy – a tool it had developed to confront the generic drug scandal – and block a number of Ranbaxy’s pending and approved ANDAs.

FDA has since gotten very good at conducting data integrity inspections, Thakur noted. In the 14 months after the Justice Department announced the Ranbaxy settlement, there were 14 instances of FDA warning letters or regulatory actions against Indian companies over data integrity failures ([A#08150130004]).

Where consent decrees won’t work

Although FDA can now see what Indian pharma companies are doing wrong, the agency will find it difficult to compel them to fix the problems, he said. That’s because he believes the weapon the agency has used so effectively to turn errant companies around in the U.S. – the consent decree – won’t work as well with Indian firms.

In India, he said, “You look at individual failure all the way to system failure. With each additional observation, you have confirmation of a system with intentional ‘holes,’ people essentially finding ways to work around the system. Is that because of lack of regulation? Is that because of lack of oversight? Is it because of a lack of training? What is it? And how do you fix this?”

U.S.-based multinationals like Johnson & Johnson’s McNeil Consumer Healthcare unit have corporate governance systems that can accept and implement the typically onerous terms of FDA consent decrees.

But pharmaceutical companies in India typically lack sufficient maturity in corporate governance, he said. The people who launched the companies retain a high degree of ownership and run them as family businesses, with family accounting for a high percentage of senior executives, investors and business partners.

Their hierarchical nature will prevent such companies from absorbing the lessons of life under FDA consent decrees, Thakur said. “There is this notion of not questioning authority. It happens in India. I was a part of that culture, and I could speak about it. The dictate comes from the top, and you have to essentially go about doing it, making it happen.”

Why firms are returning to the U.S.

Global manufacturers have been losing their enthusiasm for operating in India and other emerging economies, Thakur said, as quality issues became apparent and the economics became less favorable.

Ten years ago, the labor, electricity, natural gas and other costs of manufacturing were 30% to 40% lower in India than in the U.S., and much lower in China, he said.

Now India is at 87% of the U.S. manufacturing cost index, and China is at 96%.

“These facilities, these locations are beginning to look not that attractive anymore,” Thakur said. “We are beginning to see manufacturing coming back into the U.S.”

But the return of manufacturing won’t happen overnight, he said. “Translating this into actual action and building manufacturing facilities with continuous operation, that’s a long-term plan. That’s not something that’s going to happen tomorrow. It’s going to take 10 years to do that.”

How to strengthen network quality

In the meantime, what should pharmaceutical manufacturers do to strengthen the quality of their far-flung manufacturing networks?

The first step is to assess the risk in your supply chain, he said. “Develop a strategy to increase awareness of the risk and figure out how you can detect it early before it becomes an issue for you, especially in the R&D supply network. Remember the largest risk today is not within the confines of your own organizations. It is coming from your supply network.”

It’s important not only to assess risk when selecting vendors, but also to consider the cost of monitoring them.

And it’s important to benchmark your performance against your peers, he said. “Look at the maturity of your supply chain and sourcing compared to what your peers do, and how is that different?”

The key to choosing the right metrics

Thakur said his company, Medassurance Global Compliance Corp., remotely monitors manufacturers’ supply chains using customized sets of quality metrics. “We look at selection of certain metrics,” he explained. “We monitor those metrics and develop a dashboard for the C level.”

Medassurance’s clients are under FDA consent decrees, he noted, “so they are required to monitor this sort of information.”

Thakur evaluates the utility of metrics in Medassure’s library for individual clients in four dimensions:

Product or process quality;
Site quality;
Type of system: and
Facility category.

He would choose different metrics for a sterile injectables manufacturing facility than for one that makes oral solid doses for sale over the counter. He would typically choose balancing metrics, for example balancing lot rejection rate against right first time rate. And there would be a mix of qualitative and quantitative metrics.

He deals with issues such as determining the ideal monitoring frequency and accessing supplier information to enable real-time monitoring, “something that we’re doing right now.”

What to look for in the data

Once the data is flowing, it should be analyzed in four ways, he said:

Trend analysis, typically done by the supplier but also, to a degree, independently by the manufacturer;
Variability analysis to track challenges and improvements to process control;
Comparative analysis to view performance across manufacturing lines, sites, products, divisions and vendors, and over time; and
Forensic analysis, both structured and unstructured.

He gave examples of each type of analysis. For the forensic analysis, he showed a case where laboratory and controls had hardly any citations compared to the other areas: facilities and equipment, materials control, production, and packaging and labeling.

Normally, the laboratory incurs the most violations of any area. In fact, speaking at the same conference, FDA’s Karen Takahashi advised auditors pressed for time to start their data integrity audits at suppliers’ laboratories because of the greater likelihood of issues appearing there ([A#08141218005]).

“If you see a chart like this, the first question I ask is what’s going on here? Is it a training issue? Is it that there’s a high turnover there? Are there organization issues there? Is there data falsification? Those are the kind of questions you end up asking when you look at the outcomes of the application metrics that are used to monitor the sites,” Thakur said.

“It’s important for the manufacturers who are sourcing to pay attention to this kind of forensic analysis,” Thakur stressed. “The credit card industry does it all the time. They look at structured data and they look at patterns. … This problem has been solved before. We’re not the first people to solve this problem. There are solutions out there that will help us do this.”

Ranbaxy Whistleblower’s New Mission: How to Detect Fraud from Afar

More from India

More from Asia