Gates Package — page 965

Part of the problem is that it's quite difficult to assess the impact of many of these measures--which are broadly called "nonpharmaceutical interventions," or NPIs--in a controlled environment. Drug and vaccine trials are expensive and time-consuming (as I'll explain in later chapters), but they allow us to run experiments that test the effectiveness of drugs and vaccines. By contrast, no one's going to run an experiment that shuts down all the schools and businesses in a city just to measure the costs and benefits. Now, after two years of studying NPIs in the real world, we know a great deal about their effectiveness. COVID gave us real-world learning that no experiment ever could. At almost every level of government--city, county, state, provincial, and federal--officials have looked at data to see what was working, and hundreds and possibly thousands of academic studies have documented the impact of various NPIs. These efforts have dramatically improved our understanding of the field. The variations in policies among similar cities or countries allowed researchers to isolate the impact of individual NPIs in ways that had never been possible before. This is good news, because NPIs are our most important tool in the early days of an outbreak. Unlike the situation with vaccines and drugs, there's no lab time to study when to put mask mandates in place, figure out when to cancel big public events, or limit how many people can sit in a restaurant. (Though we will need to make sure that whatever NPIs we deploy are appropriate to the pathogen we're trying to stop.) These interventions are what we use to flatten the curve--that is, to slow down transmission so that hospitals don't get overwhelmed with patients--without having to identify everyone who's infected. If you catch it soon enough, you can find almost all the people who are infected and test everyone they have come in contact with. That's crucial, especially since it is