Generation X: Medicine's Manhattan Project: Can The World's Richest Doctor Fix Health Care?

This story appears in the September 29, 2014 issue of Forbes.

On a typically perfect summer day in Los Angeles,Patrick Soon-Shiong, the richest doctor in the history of the world, is bunkered inside his clandestine headquarters (nestled behind a security gate so unobtrusive that Uber cars consistently miss it), ready to show around a kindred spirit. T. Denny Sanford, who made a $2.8 billion fortune selling high-interest-rate MasterCards to people with bad credit scores, is now in philanthropy mode, giving away most of his fortune to children’s charities and hospitals. And he’s come to see what’s been touted as the future of medicine.

Soon-Shiong, 62, has a lot to show. First, he walks him through a mock-up of a futuristic hospital room: There’s a patch that measures a patient’s heart rate, temperature and blood pressure, and a 3-inch white cube, called an HBox, connecting every device to a computer network. He shows off a darkened room covered in computer screens: a control center from which a handful of doctors can monitor hundreds of patients, even when those patients are at home. And finally he calls up several computer programs that make sure doctors know, up to the latest scientific-journal article, the best treatment available. It’s a sweeping assemblage of data-driven toys–fueled by $1.3 billion worth of furtive acquisitions, almost entirely using Soon-Shiong’s own money.

This dizzying demonstration wows Sanford, who seems extra-rumpled next to Soon-Shiong, in his crisp tailored blue shirt and suit pants, which he fills sleekly (he owns part of the Lakers and plays hoops regularly on an indoor court at his house). “I think it’s exactly what we need in this world,” Sanford says. “I also have a hospital group. I think we’re at 40 hospitals and 150 clinics, but costs are just going crazy, and the lack of communication between these organizations is just paramount to correct.” Soon-Shiong jumps in for the close: “The hospitals aren’t organized, funded or even have the skill sets to create this kind of communications infrastructure. Frankly, the government should have done it.”

As evidenced by the incompetent ObamaCare rollout, perhaps it’s better that Soon-Shiong did, and Sanford is taking whatever this doctor prescribes. They shake hands eagerly on a deal to deploy the technology at a children’s hospital in Phoenix, Ariz.

Even after the demonstration, though, exactly what Sanford is buying remains unclear. As seen over Soon-Shiong’s shoulder, the demos look fantastic. But no outsider I spoke to had actually laid hands on all the pieces of the technology. There is no real business plan. No pricing model. All they have is Soon-Shiong’s word, which is a tricky thing. While he’s undeniably brilliant, Soon-Shiong is equally undeniably a blowhard, a view shared widely across the medical spectrum (his Twitter TWTR -1.01% handle:
@solvehealthcare). “The marketing is three years ahead of the engineering,” says John Halamka, one of the first people to ever have his genome sequenced and the chief information officer of Beth Israel Deaconess Medical Center in Boston. “What works on paper, what works in the lab and what works in a complex academic medical center are very different things.” He later adds: “Patrick is a showman of sorts, and for him to claim, ’I have solved the problems that everyone else over the last 20 years hasn’t been able to solve…’ ”

It bothers me, too. Soon-Shiong’s sparkling headquarters, a futuristic amalgamation of metal and glass where some of his 800 employees scurry about, sits in L.A.’s Culver City neighborhood, which has birthed dozens of Hollywood fantasies, including The Wizard of Oz. Accordingly, I’ve spent the past ten months trying to pull back the curtain. Soon-Shiong has allowed me an exclusive, detailed look inside his efforts–the Manhattan Project of medicine–just as he was closing the deal that will see them put into action for the first time at Providence Health & Services, a 34-hospital, not-for-profit Catholic health ministry in Oregon, California, Alaska, Washington State and Montana. And I talked to dozens of outsiders.

What was universal: the scope of Soon-Shiong’s undertaking. “When we went to see him and got a look at what they’re planning to implement, we were dazzled,” says Gillies McKenna, head of the department of oncology at Oxford University. “If you can make this work, and I agree it will be very difficult, he’s looking at an exponential increase in the amount of data we can base decisions on.” Soon-Shiong explains it this way: “We will have more information at our fingertips than we ever had in the history of mankind–every day. Not once a month, a week. Every day.” Such omniscience has the potential to reverse the perverse incentives–which emphasize treatments rather than results–driving America’s annual health costs past $3 trillion. It could also cure most of what ails us, even cancer.

Soon-Shiong is accustomed to doubters, as the son of Chinese immigrants in apartheid-era South Africa. He graduated high school at 16 and medical school at 22. His first patient, an Afrikaner, refused to be touched by him, but after Soon-Shiong drained his infected sinus, he told everyone, “That Chinaman. Make sure you get him to examine you.”

Soon-Shiong left South Africa in the late 1970s and arrived at UCLA in 1980. Stephen Nimer, a hematologist who would later serve on the board of directors for one of Soon-Shiong’s companies, remembers him as an “unbelievable surgeon” who was always willing to take on the most difficult cases. “It’s in his blood to help people,” Nimer says.

So is a deep streak of P.T. Barnum showmanship–and a talent for pissing off investors and colleagues alike. As a surgeon at UCLA he grabbed headlines transplanting insulin cells into a diabetic. The president of the American Diabetes Association called it “inappropriate hype,” saying it was “far too early to view this as a cure or even a therapy.”

In 1990 he started a company to commercialize his diabetes work and got a deal with Mylan to explore transplanting pig organs into people but backed out because he decided it might be unsafe. He ended up in a legal feud that included, among others, his own brother.

Then, in 1991 he invented the drug that made his fortune: Abraxane, which packages the top-selling cancer drug, Taxol, inside the protein albumin. The idea was that tumors would eat the albumin and get the poison.

Top oncologists called it “old wine in a new bottle.” But Soon-Shiong was convinced he was on to something big. He decided on a novel–and personally risky–approach to fund Abraxane’s development. Rather than sell stakes to venture capitalists, the traditional route to bankrolling biotech research, he instead took out loans to buy a small, publicly traded generic drug business, which he renamed American Pharmaceutical Partners, folding his Abraxane initiative inside it. A physician buying group, which purchased drugs from APP, invested in it. Some said this was a conflict of interest; Soon-Shiong says the group contributed to help prevent drug shortages and sold its shares as soon as APP went public. But his reputation had been dinged again.

In 2005 he won a huge victory: The FDA approved Abraxane, defying short-seller interest, which ran as high as 100%. Shares jumped 47%. But once again Soon-Shiong became the center of controversy when, a few months later, he merged APP with a private vehicle he owned. Brian Laegeler, an analyst at Morningstar, called it a “raw deal for minority shareholders as it serves only to line the pockets of Patrick Soon-Shiong.” The stock dropped 18% the day the deal was announced. Soon-Shiong says the long-term rise of the shares vindicated the move.

Then in 2007 the stock soared again. The firm was the only maker of the blood-thinner heparin whose product did not have to be recalled because of contamination that killed 81 people. Soon-Shiong split and sold the company, saying it was “two unique businesses.” The generics business, including heparin, went to Fresenius in 2008 for $4.6 billion. In 2010 the drug business, Abraxis, was bought by biotech giant Celgene for $4.5 billion. Soon-Shiong owned some 80% of each.

Another multibillion-dollar windfall soon followed. Despite Soon-Shiong’s insistence that Abraxane was “a breakthrough,” by 2011 sales were just $386 million–a middling success in the booming biotech sector. Then last year a study showed the drug extended the lives of pancreatic cancer patients by 1.8 months. Sales jumped 90% and are projected to hit $2 billion by 2017. Celgene’s stock–Soon-Shiong remains the largest individual shareholder–surged in lockstep.

Cleverness, determination and luck had left Soon-Shiong with enormous wealth–we put his current net worth at $12 billion. But it also left him with a reputation as more of a wheeler-dealer than a scientist, which pains him, say confidants. “He has recognition in the business community,” says Michael Crow, the president of Arizona State University, another institution Soon-Shiong is talking about working with. “But that’s very different from the recognition that this was the man who built the intellectual fabric that allowed cancer survival rates to be increased 80%.”

Soon-Shiong’s grand new project promises the closest thing that Earth has ever had to Star Trek’s fabled tricorder. In theory it will work like this: A cancer patient will arrive at the hospital for diagnosis. Everything from her DNA to the proteins in her blood will get instantly analyzed via a proprietary and superfast network, with the data collected automatically in real time–no pens, paper or clipboards. Within minutes computers will recommend which drugs to try. Once the patient is sent home, the same technology will travel with her, allowing doctors to continue to monitor her in real time, as hospital administrators evaluate the efficacy and costs of various procedures and medicines and compare notes with hospitals across the country.

This vision came during the approval process for Abraxane in 2005. Doctors were making bad decisions. One study found that two-thirds of pancreatic cancer patients received the wrong treatment. Computer brainpower wasn’t enough to fix this, Soon-Shiong realized, if it wasn’t paired with a high-tech nervous system. “How could we ever hope to win the war against cancer using our newly gained molecular insights against a disease that has the capacity to constantly change and mutate?” he says.

Like a mechanic rummaging for parts, he started buying companies to build his new machine. He grabbed Eviti, based in Philadelphia, which sold its services to insurers as a way to ensure that cancer doctors weren’t prescribing medicines improperly (and billing for their errors). Thirty oncologists and nurses pore over the latest medical journals to make sure the information is up to date.

Another purchase: iSirona, a firm in Panama City, Fla. that’s attempting to connect hospital machines with electronic health record systems. Soon-Shiong now claims that it can integrate 6,000 different medical devices, including pulse oximeters, blood pressure monitoring devices and bathroom scales, as well as hundreds of different types of clinical and financial software from every major medical vendor.

There were other technologies, too: Qi Imaging, a tool that allows CAT scans and MRIs to be viewed on mobile devices; GlowCap, an $80 pill bottle that lights up when patients at home need to take their medicine and lets doctors know they are opening the cap. He purchased and refurbished the National Lambda Rail, a high-speed government computer network, at a cost of $100 million, so all this data could move quickly from place to place. “In order to have value-based care you need to monitor outcomes in real time,” says Soon-Shiong. “And you need to monitor cost in real time. You’re going to have patient-centered highest-quality care at the lowest cost.”

All of these pieces–and dozens more that he’s bought or built–combine into a corporate structure as byzantine as his overarching product. His 800 employees are splintered across offices in 14 cities, and NantWorks, the parent holding company, houses nine separate units, all with different investor groups and each apparently designed to trade independently as a tracking stock. The first IPO, as early as next year, will likely be NantHealth, his health care information technology play, poised to profit from new payment schemes created by ObamaCare. Investors include Verizon, Celgene, BlackBerry and the Kuwait Investment Authority. FORBES values NantHealth alone at $1.6 billion. All told, FORBES values the entirety of Soon-Shiong’s Nant-related holdings at $7.7 billion.

The potential and pitfalls of Patrick Soon-Shiong’s medical Manhattan Project boil down to one statistic: 47 seconds. That’s the amount of time, the doctor/entrepreneur promises, it now takes for his amalgamated “supercomputer” to complete genomics analysis, all the way to identifying the individual protein in someone’s body that’s amenable to drug treatment. “It normally takes 11 weeks,” Soon-Shiong smiles.

Like so much he says, it’s a stunning statement with infinite promise. And an unverified one of the kind that makes Soon-Shiong polarizing.

When Soon-Shiong dramatically and fantastically described his platform in public for the first time, at the Forbes Healthcare Summit in New York last October, the preeminent doctors, scientists and health care executives in attendance ranked him the top speaker of the event (95% of attendees surveyed rated it good or excellent). Yet many of them were confused (#stilldontgetit, tweeted Halle Tecco of Rock Health, who runs an incubator for health IT companies) or skeptical (“It’s an avant-garde idea–and one that will be delivered in pieces,” said N. Anthony Coles, the former chief executive of Onyx Pharmaceuticals). Johns Hopkins professor (and Forbes.com contributor) Steven Salzberg, pondering the 47-second statement, posited: “What does that even mean?”

For the sake of clarity and credibility, I spent a disproportionate amount of my time over the past year focusing on his oft-repeated 47-second pledge. It turns out to be profoundly misleading because it is an average time, not the time for an individual. It’s like saying McDonald’s can hand you 800 Happy Meals the instant you pull up at a drive-thru window because the company serves 800 meals a second worldwide. The real question is: How fast can an individual patient’s genome be analyzed? Pushed, Soon-Shiong says he is aiming for 24 hours for each patient.

That’s still astoundingly fast. David Feinberg, the president of UCLA Health System, confirms that he got data back on cancer patients within a couple of days. Randy Axelrod, an executive vice president at Providence, where the Nant system will soon roll out, says he sent in DNA sequences from several patients and had them back within hours.

A bake-off by Genomics England, a government project that hopes to sequence 100,000 Britons, found that Nant was one of the few platforms that could consistently sequence cancer genomes quickly.

Even more impressive, Soon-Shiong says–and a number of experts believe–that he can already analyze 500 genomes a day, on par with the world’s most advanced DNA research centers, and will be able to do 4,000 a day by the end of next year. And Nant can move these huge sets of data to any hospital on his network almost instantaneously.

Which raises the question: Why the unnecessary, counterproductive hyperbole? How easy it could have been to instead take a bow for the hardware and the high-speed data connections he has built.

And why compare your time against another (11 weeks) that no longer seems to exist? “Our best shot with software that’s proprietary is 15 minutes, and then there’s still plenty of work to be done,” says Eric Topol, the chief academic officer at Scripps Health. Ultimately, whether you can do it in 47 seconds or an hour or five isn’t really what’s important. Accuracy and cost are.

Soon-Shiong angrily shrugs off the criticisms. “Unfortunately, when you go outside the accepted conventional bounds, some people feel threatened and strike out,” he says. “When you run a public company, short-sellers seize on it and give it even more legs. Fortunately, the strong persevere–and not just me–and we have a better world for it. If they didn’t, a lot of advances in health care, science and technology would never be achieved.”

Soon-Shiong’s chance to silence the doubters will soon begin. “You’ve got these fantastic ideas; you’ve got these fantastic people,” says Jim Davies, the chief technology officer of Genomics England. “Now is the bit where they roll it out.”

At St. John’s, an L.A. hospital to which Soon-Shiong has given $85 million, a prototype of a system is tracking patients’ treatments and what they cost in real time. And a full-scale deployment of Nant’s systems will roll out imminently at Providence Health & Services, which acquired St. John’s. Soon-Shiong met the system’s CEO, Rod Hochman, during the deal. They hatched the idea for using Providence not only as a testing ground for NantHealth’s software but also for its genetic tests, as it aims to offer them to every one of their 25,000 cancer patients each year.

Cancer is a disease of genetics. It happens when a genetic defect or, more likely, a collection of defects causes cells to go haywire and grow out of control. By identifying which genetic defects are present, and picking drugs to target them, doctors may be able to treat otherwise untreatable cancers.

Soon-Shiong has an example, identified using Nant technology. A woman was suffering from cervical cancer and had had her genome sequenced. When it was fed into Nant’s computers, they found that the human papilloma virus, which causes the cancer, had inserted itself into a gene called Her2. This is the target of the breast cancer drug Herceptin; when the woman was given Herceptin, a drug that would normally not be used for treating cervical cancer, her tumors shrank.

It’s a great story. But again, hype blurs the brilliance. Nant’s analysis had a wonderful result for the patient, but it’s hardly a medical breakthrough or even unique: Foundation Medicine, a cancer-genetics startup backed by Bill Gates and Google Ventures, has touted a case where cancer in a woman’s colon shrank because of a lung cancer drug.

As with everything in Big Data, what Nant brings is scale. Where Foundation Medicine tests patients’ tumors for mutations in 343 genes, Soon-Shiong plans to do 260 times more: sequencing the whole genome of the patient, the whole genome of the cancer (which is genetically distinct) and the chemical messages, known as RNA, generated by the cancer genome. Even with the use of DNA sequencers, this will cost $3,000 per patient just for sequencing, and it will take three days, plus another day for analysis. Providence expects to pay for this in part by getting insurers to pay for it.

In the end the most telling statement may have come at the Healthcare Summit from Susan Desmond-Hellmann, who watched Soon-Shiong’s rise when she ran clinical development at Genentech before becoming chancellor at the University of California, San Francisco and then CEO of the Gates Foundation: “Don’t underestimate him.”

“Deep down,” says ASU president Crow, “he knows that being the world’s richest doctor is not the check that he wants by his name. It’s insufficient.” “My quest was and is to improve the quality of life through science,” adds Soon-Shiong. “That is what drove me then, and that is what is driving me now.” The bad blood will mean nothing if he’s successful–and it seems very likely, based on a review of his claims, plans and investments, that he will succeed at something. Even a fraction of his grand vision will mean good news for millions of American patients.