DECEMBER 7, 2012
Toansa, Punjab India
As the FDA looked for investigators willing to relocate and join its sparsely staffed offices in India, the agency did not have many takers. But Peter Baker, the young consumer safety officer, raised his hand. He had a love of travel and adventure and took yearly trips with his motorcycle group to far-flung locations, from Mongolia to the Philippines. He also had a more pragmatic reason to volunteer. By reputation, India was the world leader in aseptic manufacturing, the exacting science of producing sterile drugs. Baker figured that after a stint in India reviewing best practices, he’d return to the United States with expert knowledge that would advance his career.
Just three months after arriving in New Delhi, he got a critical assignment: to inspect Ranbaxy’s Toansa plant in a remote area of northern Punjab, where the company made the active ingredient for its generic Lipitor, atorvastatin.
It was the same troubled plant that had squeaked through a last-minute inspection, largely owing to the intervention of FDA bureaucrats. But only ten months after Ranbaxy began its lucrative launch of atorvastatin, the company made a staggering admission to the FDA. It had discovered that some of the millions of pills already dispensed to patients across the United States were suffused with tiny shards of blue glass. The manufacturing failure confirmed the worst fears of the FDA’s compliance staff. Ranbaxy should never have been approved to make the drug. Now, as the company recalled millions of implicated pills, the FDA dispatched Baker to the Toansa plant, where the active ingredient had been made, to figure out what went wrong.
For the high-stakes inspection, Baker was paired with another veteran FDA investigator, Atul Agrawal, who was stationed in Mumbai. The two men were opposites: Baker was young, handsome, and athletic, with close-cut, brownish-blond hair; Agrawal was short, balding, and walked with a limp. But the two men shared one trait: they disliked being fooled. This time, because the inspection was so vital, the FDA’s India office had given Ranbaxy only a few hours of advance notice—less time to conceal evidence. But Baker and Agrawal wound up in the wrong town and had to backtrack, which took hours and bought Ranbaxy additional time.
The investigators finally arrived at the vast plant complex, with dozens of buildings stretched over acres. A corporate executive from Ranbaxy’s headquarters had beaten them there. The executive had taken the first flight from New Delhi that morning, which meant the company must have gotten word of the inspection the night before. Someone from the FDA’s India office had almost certainly leaked it.
Plant officials proposed that the investigators first sit for a presentation. Agrawal insisted that they be taken immediately to the building where the glass contamination had occurred. The officials took them to the MP-11 plant. There they explained that the contamination had come from a blue-glass protection ring that encircled the top hatch of a reactor, a big rounded vessel into which they shoveled ingredients, and mixed them to cause a chemical reaction. They claimed that they could not show the investigators the protection ring because they had since removed it; instead, they showed the men a photograph of it with a small chip in the glass. Baker studied the image of the small round imperfection, thinking that it looked too perfect. It was almost as though someone had carefully taken a hammer to it. How could that have caused enough broken glass to contaminate millions of pills?
As Agrawal and Baker inspected more of the reactors, they peered down through the hatches on top at the glass-lined walls. They even had plant employees hold their legs, as they dangled upside-down inside the reactors. As a knot of Ranbaxy officials waited silently, Baker cast a flashlight beam across the cylindrical walls of an enormous steel silo. He caught sight of a spiderweb-like crack along one of the glass linings and followed it with his flashlight. He doubted that the crack was the cause of the catastrophe. But he was certain that the hovering officials were doing everything in their power to keep him in the dark.
Inside another reactor he saw heaps of powder residue, as though it had never been cleaned. He asked to see the inspection forms. All the results were documented simply as “OK,” with no notes next to them. The cleaning checklist had been signed by two different technicians. As Baker questioned plant officials, they first spoke among themselves in Hindi, as though to get their facts straight. Amid pointed questioning, they admitted that those who’d signed the logs had not actually done the cleaning.
As the inspection continued, a plant official mentioned that the company planned to grind, refilter, and resell the glass-infused ingredients. The investigators explained that the material was too compromised, so plant officials pledged to destroy it. In the United States, it was standard protocol for investigators to witness the destruction of products deemed too dangerous to use. Baker himself had once visited a dump and watched as a company poured bleach over mountains of fish that had gone bad, before running over the piles with a bulldozer, as the fish exploded from the built-up toxic gases. Baker and Agrawal didn’t trust Ranbaxy to destroy the compromised ingredients. They later asked the director of the FDA’s India office for permission to witness the destruction. He told them it wasn’t necessary.
The investigators spent eight days at the plant. They found significant violations, but left with the sinking feeling that they’d been played—that the perfectly round hole in the glass had been faked; that they’d been diverted from a major glass failure in a reactor they hadn’t been shown. They feared that Ranbaxy would almost certainly resell the contaminated ingredients instead of destroying them, as promised.
The experience led Baker to a resolution: he vowed to get smarter and more aggressive, to pierce the coordinated defenses and denials of plant officials. His determination, and what he would soon uncover, would make him an infamous figure among Indian drug makers and raise questions around the world about their practices.
Baker grew up about as far from a world of global deceit as could be, on a grass seed farm in Lebanon, Oregon, where he was raised by Mennonite parents. His family’s church preached service to others, and the community’s missionary work took his family to remote destinations—indigenous villages down the Amazon River in Colombia and orphanages in the Dominican Republic. Baker spent summers on his neighbor’s farm, driving tractors and combines beneath an unbroken blue sky.
After attending a Christian college in San Diego, where he discovered his love of chemistry, he met a Swedish woman during a European backpacking trip. Together, they settled back in San Diego, where he got a job at a contract research laboratory that tested drugs for manufacturers. He spent hours each day running tests on high-performance liquid chromatography machines. The company seemed more interested in billing hours to its clients than in scrupulous adherence to good manufacturing practices, an attitude that trickled down to the lab. There, minimally supervised chemists used shortcuts to nudge data up or down in order to finish their work quickly, particularly on Friday afternoons when the company brought in a keg and the employees played basketball.
After a year, Baker and his girlfriend moved to China, where he taught English and Western culture to engineering students. On their return to San Diego, they married. Baker wound up at Abbott Laboratories, working as an engineer at the company’s quality assurance laboratory. His real learning began there. His boss assigned him the task of taking a lab with no formal controls and bringing it up to code. This meant creating an entire infrastructure of transparent and verifiable standards. Baker learned GMP from the ground up.
In 2008, while completing a master’s degree in analytical chemistry at San Francisco State University, he joined the FDA. Within a year, he was inspecting plants in the United States. The facilities were largely compliant, and he’d spend weeks at each one, urging companies to perfect the manufacturing process. Only two of his fifty-five U.S. inspections resulted in a warning letter. Afterward, one of the companies ended up moving manufacturing plants to Taiwan, in part to evade so much scrutiny. By mid-2011, with the FDA facing a backlog overseas, Baker volunteered to do some inspections in China, planning to make use of the primitive Mandarin he’d studied in college.
Months later, Baker was asleep in a hotel in Pucheng, a remote town in Fujian province, where he had traveled to inspect a veterinary medicine plant. He awoke at 1:00 a.m. to pounding on his hotel door. Through the peephole, he saw more than a dozen men in street clothes, who turned out to be part of a government security team. Before he could unlatch the door, they busted it open, swarmed in, briefly detained him, and searched his room. The hotel later claimed to have confused him with someone else—though he was almost certainly the only foreigner at the hotel.
The FDA had given him only cursory training for the world of foreign inspections he had entered, which included the admonishment not to accept any gift worth more than $20. Topics like government surveillance, offers of gifts that included gold coins or call girls who materialized in hotel rooms late at night, and how to stay safe in remote towns controlled by the companies whose plants he would be inspecting never came up. As Baker learned to navigate the dicey situations, he found that the long hours, forlorn towns, and unpredictable nature of the inspections energized rather than daunted him. Within months, he agreed to relocate full-time to India.
In late 2008, the FDA opened several foreign outposts, including in India. The move was a response to the adulterated heparin crisis and an acknowledgment of what had long been obvious: the FDA couldn’t hope to regulate plants in a country where it had no permanent presence. But the FDA’s India office had gotten off to a rickety start, with a skeletal crew of eight employees—four in Mumbai and four in New Delhi—tasked with inspecting hundreds of plants across the entire country. Investigator Mike Gavini, who’d been so lenient with Indian companies on his visits there, was among the first to volunteer. He joined the Mumbai group in June 2009.
In Mumbai, the investigators worked out of the building that housed the U.S. Consulate General, where mice chewed through the power cables on their computers. Gavini was housed miles away and commuted on a halting local train. He traveled to manufacturing plants that had never seen an FDA investigator before. He found untrained and uneducated workers, some of them illiterate, traipsing through factories in street shoes. But in his effort not to “indiscriminately [shame]” the country people, as he would later put it, he nudged, cajoled, and guided India’s reluctant factory owners, often giving them green lights to proceed based on their promises to do better. As the companies thrived under his lenient inspections, the number of Indian plants approved to sell drug products to the United States soared. By the time Baker arrived in September 2012, India had become a pharmaceutical-exporting juggernaut.
As Baker started his inspections, he saw what had given India its vaunted reputation: many of the manufacturing plants were new and appeared to be in immaculate condition. The equipment was pristine. But he noticed something else. From the minute the inspections began to the minute they ended, the managers led the FDA investigators around the plants like dogs on a leash. First, a welcome reception. Then, an opening slide show. Guided tours of the facilities. The elaborate show of hospitality amounted to running down the clock.
On his fifth inspection in India, at RPG Life Sciences in Mumbai, Baker took a detour and visited the company’s quality control laboratory, an out-of-the way spot where the plant aggregated its testing data. There, he asked to see all the records related to failing drug products. A nervous manager admitted that the records had been destroyed. This troubled Baker. But as he scrolled through the company’s computer systems, he found some documentation about the products that had not been officially logged. When everything was supposed to be transparently documented, why did the company have an unofficial set of records? He gave the plant a flunking grade of Official Action Indicated. In safe, well-run plants like those he had inspected in the United States, there were not supposed to be any hidden or destroyed records, back doors, or unofficial anything.
In January 2013, his fourth month in India, Baker arrived at a manufacturing plant in Kalyani that made active ingredients for injectable cancer drugs. It was owned by Fresenius Kabi, a German brand-name drug company. The last FDA inspection there four years earlier had only one finding: the company had not correctly documented the chain of custody for the drug samples it was testing. Since his experience at the Toansa plant a month earlier, Baker vowed that he would no longer allow himself to be steered through a manufacturing site. Instead, he would choose where to go and what to inspect. As at RPG, he headed directly to the quality control laboratory—the ultimate back room.
The ostensible purpose of a quality control lab was to audit the test results coming in from the factory floor to ensure that they were unaltered and preserved and that any irregular data was investigated. It was also a last line of defense for a plant interested in detecting, and eliminating, failing drug products.
But the labs could also play a nefarious role, as hubs for manipulating or discarding data that might expose those failing drugs. At the Kalyani lab, instead of asking for data, Baker simply sat down at a computer and began to scrutinize the results from tests done on high-performance liquid chromatography (HPLC) machines. The bulky machines separated and measured the impurities in a drug sample and displayed them as a series of peaks in a record called a chromatogram.
Toggling back and forth between computer files, Baker discovered something strange. He found the official test, with the result stored in the correct data folder. But in a file called “MISC.,” he saw what looked like earlier tests of the same drug sample, some a day apart, some a month apart. Some of the earlier tests were in ancillary folders, like “MISC.,” or within the proper folder but labeled “DEMO.” This unofficial pretesting and the subsequent retesting were not explained. Nor were there any procedures that permitted it. He figured out that some of the HPLC machines were not officially registered and had not been connected to the plant’s main server. The plant was conducting offline tests.
Baker had uncovered the outlines of a secondary manufacturing operation hidden beneath the surface of the first. The technicians were using the initial hidden tests to get preliminary results, which they used as a guide to tinker with the test settings—adjusting the parameters, the amount of solvent. Then they retested in the plant’s official system to get the desired results. As Baker showed the plant managers point-blank evidence of this malfeasance, they denied his findings. All test results were stored on a central server, one manager insisted, despite all the data Baker put in front of him. As Baker’s inspection progressed, he moved deeper into the plant. He opened an unmarked binder and found scratch paper. A manager snatched it from him and tried to hide it in his pocket. Baker demanded it back. The paper contained notes about manufacturing errors and other problems that needed to be fixed. The hidden notes clearly sidestepped the plant’s obligation to document all quality issues in official records. Baker also noticed foreign particles in a closed vessel used to filter solvents, which the manager denied, claiming it was due to the reflection of light on the glass. When he returned later, the particles had been removed. The manager denied this, asserting that the technicians would not have undertaken “cover-up activities.”
The next day Baker unearthed a report that confirmed his suspicions: it stated that operators had removed the “particulate matter” during his one-hour absence from the processing room. The fragments had come from a deteriorating internal gasket. A statement from the manager was attached: “During the USFDA inspection on 14th January 2013, I panicked and advised my junior to open the equipment and clean it.”
As Baker asked questions, plant managers in India were relaying these to quality managers in Germany, sparking alarm. On the last day of the inspection, Baker faced a stunned and exhausted executive vice president, who’d flown all night from Germany with an entire team to be there when Baker rendered his final verdict. It was evident that the Germans hadn’t known what was happening in their own plant. Baker’s questions had caused them to grill their employees there.
The Germans were forthcoming with Baker and shared what they’d learned in the prior seventy-two hours. The plant’s managing director had overseen a scheme to pretest drug ingredients, preview the results, make secret adjustments to the test settings, and then retest the drug samples until they passed. The director had ordered the HPLC machines on which they’d engineered the pretesting shipped offsite before Baker’s arrival. The plant had reblended failing drug ingredients that had high impurities with higher-quality ingredients until they passed testing. Baker was particularly stunned to learn that the most senior person at the plant had directed these activities.
In the months that followed, Fresenius Kabi launched a probe into its own plant. The findings were grim. The company later admitted to the FDA that it had found unofficial testing, unauthorized blending, false manufacturing records, and the deletion of thousands of records, and that it could not trust any data from the plant, or the compliance of any batches, prior to Baker’s inspection. It was a debacle. To its credit, Fresenius Kabi halted production at the plant, fired the entire management, and recalled all the medicine made with reblended ingredients. As shocking as the situation was, Fresenius Kabi seemed to be a victim of its own poorly managed outsourcing. Had Baker not uncovered the malfeasance, the company might never have discovered what was happening in its own plant.
But Fresenius Kabi wasn’t alone. Brand-name and generic drug companies the world over were snapping up plants in India to make their own active ingredients and drugs at a fraction of the cost. They could rapidly increase their profits, all from saving on the cost of labor and supplies. But Baker suspected that company owners had little insight into what really went on in their own money-saving plants.
Before Baker arrived in India, the FDA’s investigators had long suspected that something was not right. Typically, a compliant manufacturing plant will reject a certain percentage of drug batches for many different reasons. But in India, investigators rarely saw a rejected batch. Somehow, almost all of them passed. Plants were also missing documentation. “I’ve always known [Indian companies] had a hard time keeping paperwork,” one FDA investigator explained. “Their habit and practice is not to keep documentation. They have a chalta-hai attitude,” he said, using a popular phrase in India usually accompanied by a shrug, which means a willingness to accept a less-than-ideal outcome. “In my mind, I knew something had to be not right.”
But with advance notice and only a week’s time at a sprawling plant complex, it was hard to know precisely what was not right. Baker changed that, in part by looking in places where no one else had. As he scoured data, he followed clues: audit trails within the supposedly impermeable software systems had been disabled; tests had been repeated, and then deleted from the official network servers. Through a painstaking forensic effort, he matched up the metadata from deleted tests with the later official tests of the same drug samples. Among hundreds of thousands of tests at any given plant, he followed his instincts: he zeroed in on drug tests he suspected the facility might be faking, and the equipment they might be using to fake them.
Baker became good at finding the HPLC machines that were used for the secret tests and not linked up to a plant’s central software system. In some cases, the company hid them in concealed labs, or even amid other networked machines. But Baker began finding them—largely because he began looking for them. Previously, investigators walked into a laboratory with dozens of HPLC machines and allowed company officials to guide them to a machine, to show them how it worked and what data was being generated. With only five days at a facility and so much to inspect, finding the one or two non-networked machines was nearly impossible. During one inspection, he simply asked lab workers, “Are all these units networked?” One of the lab technicians piped up, “That one is not.” A regional manager tried to cut him off: “Yes, it is.” “No, it is not,” said the employee. Back and forth they went. But Baker had uncovered a stand-alone machine simply by asking for it.
As he worked, other FDA investigators learned from his techniques. The difference was dramatic. “It was like ‘holy shit,’” one FDA employee recalled. “Like you walk into a dark room and suddenly someone just turns on the light. It was shocking.” But it was a fateful inspection two months later—when he discovered the torn batch records in a garbage bag that an employee tried to slip from the plant—that darkened Baker’s view considerably and focused his thinking about the dangers of faked data.
On March 18, 2013, Baker arrived at the main plant of Wockhardt Ltd. in the Waluj area of Aurangabad, two hundred miles east of Mumbai. At the inspection’s opening meeting, the company’s vice president of manufacturing insisted repeatedly to Baker that the plant had only one manufacturing line that it used to make products for the U.S. market. But on the inspection’s second day, Baker recovered a garbage bag that an employee had hurled beneath a stairwell. Inside were torn batch records for the company’s insulin products, which showed that numerous vials contained black particles and had failed visual inspection. As Baker soon discovered, the test results had not been logged into the company’s official system. The manufacturing line on which the company had made the drugs didn’t exist in official records but operated secretly within the plant. In official records, the insulin products passed inspection at a much higher rate than in the deleted batch records. The drugs had been released into Indian and Middle Eastern markets.
Though Baker’s mandate only extended to drugs for the U.S. market, he and his colleague, a microbiologist, continued to follow the clues in the torn records. The next day they found the secret formulation area and unearthed an unofficial “investigation report” into the black particles. The report was not dated or signed but was written to the head of the plant. It made clear that the black particles were “metallic” and had come from damaged heating coils inside a machine used to sterilize vials and cartridges before they were filled with medicine. The blistering heat inside the machine had broken down the coils. They had been repaired, but hadn’t been replaced due to cost. The medicine with the particles had been released to patients.
Baker was staggered. It would have been one thing for the company not to know about the particles, or to accidentally order the drugs released. That would have been shoddy or negligent. But the plant director had ordered the insulin released, with full knowledge that the drugs were potentially lethal to patients. The metallic particles could easily cause an immune-compromised patient to suffer anaphylactic shock and die. Worse yet, Baker discovered that an injectable drug for the U.S. market, adenosine, used to treat an irregular heartbeat, was manufactured on the same secret line, and with the same perilous equipment. Though the unofficial report noted nothing about that drug, Baker was sure the results were no better.
If he’d found such deceit in the United States, it would have triggered a law enforcement raid and, almost certainly, prosecution. Someone would be going to jail. But in India, Baker had no such authority. The only remedy was a regulatory one. He had enough evidence to restrict all the drugs made at Wockhardt’s Waluj plant from entering the U.S. market.
The inspection lasted five days. Baker’s findings—seven major ones—were devastating. They painted a picture of elaborate fraud, extreme hazard, and filth. Plant officials had repeatedly refused to cooperate. One official dumped vials into a drainage sink after Baker inquired about their contents. In the bathroom, about twenty feet from the gowning area for the sterile formulation lab, the urinals lacked drainage piping. As Baker noted in his report, “Urine was found to fall directly onto the floor, where it was collected in an open drain.” The room had an “overwhelming sewer stench,” he wrote. By mid-inspection, both Baker and his colleague had fallen ill, after plant officials gave them unsealed water bottles. They suspected that plant officials had tried to sicken them, to shorten their inspection time.
Word of Baker’s findings, let alone his published inspection report, could wipe millions off the company’s stock. At the inspection closeout meeting, a part of agency protocol, Baker presented his findings to the company’s vice president of manufacturing. The executive demanded, with a menacing glare, that Baker remove the first observation, about the discrepancy between unofficial and official insulin batch records, from his findings. It was a threat.
“I can’t do that, sorry,” Baker responded, getting increasingly nervous about safe passage from the plant as the man stared him down. “Let’s get out of here,” he told his colleague. Neither wanted to get into a company car. The firm was in the middle of nowhere, and a deadly traffic accident would surprise no one, given the number of trucks in the area and the craziness of the roads.
His colleague wanted to mail back the evidence she’d collected rather than travel with it. The company volunteered to call DHL. The next thing they knew, a man dressed in what appeared to be a fake DHL uniform wandered in. A plant employee clearly intended to take her exhibits and disrupt their inspection efforts by any means possible. When Baker demanded to see the man’s DHL van, he left and didn’t return. The FDA investigators were spooked and headed out to the road with all their gear to hail a tuk-tuk, a three-wheeled makeshift cab.
Baker noted in his inspection report: “Due to the threatening behavior and personal safety concerns encountered during this inspection, it is suggested that an inspectional team perform the follow-up inspection with a clear emergency plan in place prior to arrival.”
For Baker, the Wockhardt inspection of March 2013 was a watershed. Amid the hidden laboratories, secretly repeated tests, and altered results, Baker realized that he’d unearthed far more than individual acts of fraud. He had uncovered the larger game being played by India’s generic drug industry: making third-world drugs and selling them at first-world prices. The companies exploited their technological savvy, captive employees and a corporate culture inured to fraud. They were also aided by the FDA’s archaic inspection methods and the West’s dependence on cheap drugs.
Americans had embraced the promise of cheap generics: identical drugs for a fraction of the cost. They assumed that faraway companies were making them with no compromise to quality, in large part because the FDA claimed it was so. But the Indian companies Baker had caught were aiming to make the lowest-quality drugs they could get away with, to make the biggest profits. Without a doubt, the companies could have made perfect medicine. There was no knowledge gap. Their equipment was first-rate. The difference was simply cost. Exacting controls cost about 25 percent more, according to some industry estimates.
To avoid significant up-front costs without a guarantee that their drugs would be approved, the companies pretested everything in secret labs tucked within their manufacturing plants. There, they screened in advance for failing results in order to make secret adjustments to tests. They worked to make their formulations look perfect on paper, regardless of their actual quality. They jury-rigged the results by fiddling with the tests, retesting already proven batches, or even testing the brand-name products instead. Only then would they move the data to the computer system the FDA was going to examine. Until Baker showed up in India, only Ranbaxy had gotten caught—because of the whistleblower Dinesh Thakur.
Baker uncovered all this by breaking the mold of traditional inspections. Instead of roaming the factory floor and checking equipment maintenance records, he focused instead on forensic analysis of the companies’ computer systems, an effort in which he had no formal training. He’d taught himself. It was a risky approach. If he’d spent an entire week combing through computer files and emerged empty-handed, he’d have little time left for the more conventional inspections he was required to do. But Baker had figured out what to look for: in files like “MISC,” “CHRON,” and “DEFAULT” he found tens of thousands of secret tests not entered into the quality systems.
This bogus manufacturing system required the knowledge and participation of hundreds of people, which would be nearly impossible in the United States, where plants faced the constant threat of unannounced inspections and employees had access to more responsive regulators and a chance of becoming legally protected whistleblowers. But in India, if employees challenged the practices, they could face banishment from the industry, if not worse. In order to get a new job, they needed a letter of recommendation from their previous employer. Survival dictated that they leave on good terms. Becoming a whistleblower could be lethal. It was a perfect system for fraudulent manufacturers.
Many of the employees at generic pharmaceutical plants were contract laborers, and at remote plants, some were even poor farmers. They were poorly trained, if at all, and often illiterate, despite the requirement that they document and sign off on activity logs during their shifts. By regulation, they had to be regularly tested on the rules they had to follow. One plant kept all the answers to a test up on the wall, so that the workers just had to look and copy. Most of the employees lived on one meal a day. In their daily lives, many lacked access to a toilet or running water. To Baker, it seemed ludicrous to expect them to walk into a sterile manufacturing plant and suddenly follow all the rules.
Over time Baker refined his inspectional methods, getting better and sharper and quicker at finding hidden fraud, whether concealed in ancillary files, deleted from computer systems, tossed in the trash—or even shipped out of the factory before his arrival. Fear spurred him on. If there was fraud and he failed to find it, no one else would. The finished doses manufactured at Wockhardt and other Indian plants were shipped straight from those factories to American wholesalers and drugstores. People had the right to know what they were taking and to choose what they didn’t want to take. But American patients had no idea of the subterfuge that went into the manufacturing of their low-cost medicine, and the FDA had no plans to tell them. He felt that he was the last person standing between the American patient and companies determined to cheat when given the chance.