Yesterday’s drug-research-is-hard screed brought a lot of responses, and I wanted to address some particular points that came up. On Twitter, Alex Tabarrok of Marginal Revolution advanced the Hollywood Analogy. I’ve used this one myself; it leads off a talk that I put together a few years ago. The model for a movie studio is that not all your films are going to be hits, or even make money. You can make and release a certain number of them per year, with fairly long lead times, and you try to arrange those to give you the best chance of success. No one’s really sure what makes a bit hit, though, so there’s a lot of uncertainty.
So far, so pretty good. That really does sound like the drug industry. The biggest difference (as I said in that talk of mine) is that Hollywood has no FDA. No one’s going to sue you if they weren’t entertained by your movie; the government isn’t going to make you withdraw it from distribution if it wasn’t funny or exciting enough. (Neither is the government going to examine your movie in detail before release, asking you to substantiate its billing as “The Feel-Good Hit of the Summer!” or change the labeling).
This current response, then, is to take that objection and run with it. Instead of imagining the movie industry as if it had the regulatory environment of pharma, what if pharma had the regulatory oversight of Hollywood? What if companies just released their proposed therapies (after, one presumes, some safety work) and the consumers just decided where to spend their own money, the same way that they do at the theater?
I think that the best way that I can deal with this analogy is to take on the various ways in which it breaks down. These are not necessarily in order of importance – not everyone will agree on that order – but here we go. First, we have a big one: we know a lot more about movie-making than we know about drug-making. And by movie-making, I don’t mean only cinematography, acting, set design and all the physical processes that go into the process. I mean narrative structure and the forms of entertainment. I won’t go into all those lists of the basic plots, or into a discussion of common screenplay structures (and how they’re tweaked or played against for effect). I just want to note that these things exist, even those the great majority of people watching the movies don’t notice them much.
That’s why movies can be pitched in single lines to people who know a lot about them: “Star Wars, but underwater”. “Treasure of the Sierra Madre, in Vietnam”. “Rom-Com, with a twist in the third act”. These days, they’re most pitched in terms of sequels and prequels, making the explanations even more telegraphic. My point is that narrative entertainment is a very well worked out area. Heroes go on journeys, plucky young outcasts fight against the odds, lovers discover after misunderstandings and mistakes that they were actually right for each other all along. An outsider travels along observing the ways of the natives through fresh eyes, a trusted friend and partner turns out to have betrayed the team, a sudden tragedy changes everyone’s lives. Do any of these sound familiar at all? These are tropes, and if you want to see them elaborated at the wholesale level, read one of the books that are pitched at people who want to write novels or screenplays. For the retail experience, go to tvtropes.org, a notorious time sink whose community of obsessives has broken these things down into insanely small granules, copiously annotated.
The equivalent in drug research is the idea of looking for agonists, antagonists, or allosteric modulators at receptors, or inhibitors of enzymes. These are our well-worn plots, but there’s a big difference, in that we’re not trying to please humans by doing what humans have liked before. We can’t pitch a new enzyme inhibitor project by saying “This is like captopril, but without the sulfur”, because those lessons just don’t carry over. Every project is new. We learn some general lessons, but the number of exceptions and idiosyncracies always threaten to swamp them. Fundamentally, the problem is that we don’t have control over the story in the way that movie-makers do. We have to adapt to the reality of what’s taking place inside the human body, which was not designed for humans to tinker with. This is another variant of what I’ve called the Andy Grove fallacy, where Silicon Valley/computer science types see drug research through their own experiences in coding and hardware development. In both cases, people are confusing the relative ease of working in a human-produced environment for the difficulties of working in one that has no regard for us at all.
That’s a big one. A second big difference is that the stakes are higher. If you blow $10 on a movie that you didn’t like, that’s not a fun way to spend a couple of hours. But if your new drug doesn’t help your disease, you may well be far more upset – and even more so if it turns out to make you even sicker. I’ve talked about how health changes people’s opinions about their economic behavior here, and many other people have gone into this topic at greater depth. Basically, your own health is the biggest ace in the deck – without it, little else can seem worthwhile. Comparing it with a choice in movies gets stranger and stranger the harder you look it it.
Third, there’s also a time-and-choice problem that does not occur in entertainment. Many diseases (acute and chronic) are easier to treat when caught early, which makes an incorrect diagnosis or therapy matter a lot. There may well not be a chance to make it right – you can’t go back next weekend and see a better movie or rent a version of the story that you’d like better. That window has closed.
Fourth, that time-and-choice difference applies on the production end as well. Few movies have their principal photography schedule last for four, six, or ten years (well, maybe Orson Welles’ last movie, but you know what I mean). Drug project routinely go on for just those spans, tying up greater and greater amounts of capital along the way. A movie has to be cast, shot, edited, and distributed, but the risk profile through that process is much different than it is in drug research. The latter stages of a drug development effort are by far the most expensive (Phase II and especially Phase III trials); it’s as if getting the final edit of a movie consumed 70% of the total budget, and as if the editing technology would (in a majority of cases) suddenly and randomly zap your entire film into unrecoverable waste products. Hollywood would look a bit different under those conditions.
Fifth, another reason that Hollywood differs greatly from drug research is copyright versus patent law. This is so obvious that it shouldn’t need belaboring, but in the US, such works of art are protected for at least 95 years after release. Drug patents are for 20 years after filing, and the drug itself will not appear for some years after that filing date, if it ever does. Huge best-selling drugs go generic all the time, but movies? In 2019, US films made in 1923 will start to enter the public domain. Hollywood, again, would run rather differently if cheap copies of (say) all the Disney movies and characters could be sold after twenty years.
A sixth difference is in the ways that art and science march along. Aspirin was first synthesized in 1853, and first sold in the late 1890s. Nitroglycerine was first used as a drug in 1878. Ibuprofen’s patent was filed in 1961. These are all still perfectly good drugs, and in everyday use. At the same time, an older drug or treatment can be totally wiped off the medical map by something definitively better – no one revisits them by taking them out of nostalgia or historical interest, as someone would watch an old film. Plenty of drugs from the 1970s, 1980s, and 1990s fall into these categories – while some have been superseded, others are going to be with us for a long time to come. They don’t look out of fashion; their colors have not faded, they don’t have odd slang in their dialog, and their special effects are not dated. We’re on a far different time scale in the drug and medical world, because it’s again not a human one. The English language will change long before the cyclooxygenase enzymes do.
So while the Hollywood analogy is still one of the closer ones to drug research, on the absolute scale, it’s not all that close at all. The superficial resemblances are actually a bit of a trap, leading people to extend lines of reasoning further than can be supported. I hate to keep banging away on this theme, but drug discovery is different, different, different. No other industry works this way.