Revenge of the Tipping Point: Overstories, Superspreaders, and the Rise of Social Engineering

by Malcolm Gladwell

If the world can be moved by just the slightest push, then the person who knows where and when to push has real power. So who are those people? What are their intentions? What techniques are they using? In the world of law enforcement, the word forensic refers to an investigation of the origins and scope of a criminal act: “reasons, culprits, and consequences.” Revenge of the Tipping Point is an attempt to do a forensic investigation of social epidemics.

That’s the revenge of the Tipping Point: The very same tools we use to build a better world can also be used against us.

In epidemiology there is a term called the “index case,” which refers to the person who kicks off an epidemic.

An epidemic, by definition, is a contagious phenomenon that does not respect borders.

How your doctor treats you, in many cases, has less to do with where your doctor was trained, or how well he or she did in medical school, or what kind of personality your doctor has, than with where your doctor lives.

Wennberg and other researchers have found that small-area variation does not result from what patients want their doctors to do. It stems from what doctors want to do to their patients.

What he found instead were medical clusters, where the doctors in one hospital district took on a common identity, as if they had all been infected by the same contagious idea.

Whatever contagious belief unites the people in those instances has the discipline to stop at the borders of their community. There must be a set of rules, buried somewhere below the surface.

If epidemics are influenced by the overstories created by the inhabitants of a community, then in what sense are communities responsible for the fevers and contagions that plague them?

The best solution to a monoculture epidemic is to break up the monoculture.

Centola ran this game over and over again, and always got the same result. The majority’s consensus fell apart when the number of outsiders reached 25 percent.

Yet the idea that there is a magic moment somewhere between a quarter and a third is different. It practically begs us to intervene.

When a group of educational researchers led by Tara Yosso looked just at classrooms where the percentage of minority students exceeded 25 percent, they found the test-score gap completely vanished.4 The white students did as well as they always did. But now the black students had caught up.

The game that Harvard is playing with rugby and that Georgetown is playing with tennis is, of course, also affirmative action. Except that instead of admitting underprivileged students with lower academic credentials, athletic affirmative action admits privileged students with lower academic credentials.

The United States decided that it had no place for a special deal intended to benefit people who have been subjected to discrimination and hardship. But it was fine with a special deal intended to benefit people who can afford to spend hundreds of thousands of dollars on their children’s ground strokes.

The final estimate was that the Biogen meeting resulted in well over 300,000 infections. And how did it start? “We assume,” Lemieux said, “it was introduced by one person.” More than 300,000 infections from a single meeting, all traced back to a single person.

“That’s what we call superemitters,” Ristenpart said. “Some people just released about an order of magnitude more aerosols for the… same given observed loudness.” He went on, “We had no idea. If I had to go back to the beginning, I probably would’ve hypothesized: Different people have different-sized distributions. But I didn’t guess it would be an order-of-magnitude difference between people.”

The British study, by contrast, was the first to look at what happened to people with COVID. And what did they find? A full 86 percent of all of the COVID virus particles detected in their group of infected volunteers came from… two people.

“For unclear reasons, certain individuals are ‘speech superemitters’ who emit an order-of-magnitude more aerosol particles than average,” Ristenpart and his colleagues wrote in their Aerosol Science and Technology manifesto.

William Ristenpart thinks that superspreaders might be people who, by some quirk, have saliva with unusual properties: Their saliva is more elastic and more viscous—thicker and stickier—than normal.

David Edwards, for his part, believes that whatever individual differences exist—at least when it comes to particles emitted by breathing—might be amplified by something as simple as hydration.

“Your upper airways are like a car wash,” he explains, “and the air that comes into your upper airways is like a car.” When the car wash is working properly, the vast majority of the little bits and pieces in the air you breathe gets washed away. “If you stay well hydrated, your upper airways will capture pathogens all the time, and they move them—within twenty minutes or an hour—out into your gut and you swallow… and they’re eliminated that way,” Edwards said. “But when you’re dehydrated, there’s no water in the car wash.” And with the car wash broken, things like virus particles get past the cleansing operation in your upper airways and into your lungs. That’s why being dehydrated makes you more vulnerable to colds and the flu and COVID: When you exhale, those virus particles come back out—and now you are more likely not just to contract a virus but to spread it. The particles hit your dry airways and break up into a concentrated, foamy spray, like a big wave hitting a beach. That’s how you get to 3,545 particles per liter.

The older you get, the more dehydrated you tend to be. The more body mass you have, the more dehydrated you tend to be. And when you become infected [with COVID-19], you often become dehydrated. And so it turns out that the common denominator in these three groups of people is a dehydration issue.

We don’t yet know which—if any—of these explanations is correct.

The triplicate laws seem to have a dramatic effect on the product usage behavior of the physicians. Specifically, the groups in Texas revealed almost no use of the Class II narcotics for treatment of non-cancer pain.

The [primary care physicians] and surgeons in the non-triplicate state (New Jersey) indicated a very high likelihood of using OxyContin for selective treatment of non-cancer related pain, and the rheumatologists in Connecticut also felt it had a place in their practice. However, the doctors in the triplicate state were not enthusiastic about the product at all…

our research suggests the absolute number of prescriptions they would write each year is very small, and probably would not be sufficient to justify any separate marketing effort.

The launch of OxyContin—one of the most sophisticated and aggressive drug-marketing campaigns the world of medicine has ever seen—was targeted at states without triplicate laws. So, no big push in New York State. But yes to West Virginia. No to Illinois. But yes to Indiana. No to California. But yes to Nevada. No to Texas and Idaho. But yes to Oklahoma and Tennessee—with the result that the opioid epidemic did not hit the entire United States equally.

Let’s look again at the top five opioid-consumers. These are all “Portenoy states,” without a triplicate program. Nevada 1,019.9 West Virginia 1,011.6 Tennessee 938.3 Oklahoma 884.9 Delaware 881.5 Here is the per-capita opioid consumption for the same year in the Madden states. Illinois 366 New York 441.6 California 450.2 Texas 453.1 Idaho 561.1 Illinois had one-third the opioid use of Nevada and West Virginia. New York had half the problem Tennessee did. Of the triplicate states, only Idaho came anywhere close to the national average.

If New York had Massachusetts’s opioid-overdose rate from 2000 to 2019, the economist Abby Alpert estimates that an additional 27,000 New Yorkers would have died of overdoses. 27,000. Massachusetts is not poorer than New York. It doesn’t have a higher unemployment rate. It doesn’t have a bigger problem with gangs or organized crime or drug trafficking. The two states are like peas in a pod. The only relevant difference is that half a century ago New York forced doctors to make two additional carbon copies of every prescription they wrote—and Massachusetts did not. And those carbon copies saved thousands of lives.

Here is the economist Yongbo Sim’s conclusion, after comparing crime rates in triplicate and non-triplicate states: “I find that non-triplicate states at the time of OxyContin’s introduction experienced a relative rise in both property (12%) and violent (25%) crimes compared to states with the triplicate prescription policy (triplicate states).”

Between 2007 and 2016, the number of doctor visits by OxyContin sales reps goes up by nearly a factor of five. And that’s not to every doctor in Tennessee; that fivefold increase was directed at the superspreaders.

In Chapter 6 we discussed the fact that superspreaders are profoundly different from the rest of us: There is something inherent in their physiology that makes them capable of producing orders of magnitude more virus than everyone else. The same was true of their superspreaders, Purdue discovered: They weren’t wired like most doctors. When the Purdue sales representatives downplayed the risks of addiction—making the nonsensical claim that the long, slow, smooth pace with which the drug entered the bloodstream meant that it produced none of the euphoric highs that led to dependency—the superspreader believed them. When it became clear that OxyContin was being abused—that people were just grinding up OxyContin tablets and snorting the powder, getting twelve hours’ worth of opioid in a single giddy rush—the superspreader was indifferent, or oblivious. The superspreader thought handing out drugs willy-nilly was what a doctor did. One of Purdue’s targets in Tennessee was a doctor named Michael Rhodes, who ran a pain clinic just north of Nashville.

Unlike most doctors, the Super Core did not like to be ignored by salespeople. If you visited them one to four times a year, their prescriptions still declined. Even if you visited them eight times a year, twelve times a year, sixteen times a year, they still fell. The Super Core wanted love, and that frequency simply wasn’t enough. But if you visited them twice a month—month in, month out—what happened? The number of prescriptions they wrote leaped. Twenty-four visits a year was the tipping point. If you held the Super Core’s hand and wined and dined them, they would be your best friend forever.

Between January 2010 and May 2018, someone from the Purdue team visited “the candyman” 300 times. In the past eight years, have you seen even your best friend 300 times? Once the opioid epidemic was off and running, the epidemiologist Mathew Kiang calculated, the top 1 percent of doctors “accounted for 49 percent of all opioid doses.” People like “the candyman” and Michael Rhodes prescribed 1,000 times more opioid doses than the average doctor. Purdue fueled an epidemic that would end up consuming the lives of hundreds of thousands of Americans based on the seduction of no more than a few thousand doctors concentrated in a handful of states.

The opioid crisis unfolded in three acts. The first was the decision by Purdue to avoid the states that subscribed to the Madden overstory. The second act began with McKinsey’s diabolical reinterpretation of the Law of the Few. But the third act was perhaps the most catastrophic. It was when the group proportions of the crisis changed.

The final chapter in the opioid crisis began without fanfare. In the summer of 2010, Purdue made a terse announcement. The old OxyContin was to be retired. It would be replaced with what the company called OxyContin OP. OP looked the same. It had the same ingredients. But unlike the previous version, it couldn’t be crushed into powder and snorted.2 It had the consistency of a gummy bear. The days when an addict could grind up one of Purdue’s pills and get twelve hours’ worth of opioid in one quick hit were over.

We would have been better off had we said no to Purdue’s reformulation in 2010 and kept things the way they were. But how on earth would we have done that?

Someone would have had to stand up back in 2010 and say, Look. We have two versions of a highly addictive drug. The original version is easily abused. The new and improved version is not. But we don’t want the new and improved version. We want people to continue to crush their OxyContin and snort it, the way they have for the last fifteen years. Can you imagine the reaction if health officials had taken this position? “It would be a really wild policy prescription, right?” Powell said. “It’s the craziest idea. But I mean, given what we know now, I think it’s the right one to have made. Yeah. You’d do it in a heartbeat.”

The tools necessary to control an epidemic are sitting on the table, right in front of us. We can let the unscrupulous take them. Or we can pick them up ourselves, and use them to build a better world.