Last week, the FDA approved a new drug that can cure Hepatitis C. One pill costs $1,125.

This is a bonanza for Gilead Sciences, the manufacturer of the drug, but is cause for concern among Medicare or insurance companies, who will have to pick up the tab – unless they decline to, in which case patients are on the hook (estimated annual cost: $90,000/patient). It’s a perfect illustration of how medications can do a great deal of good — but also cost a great deal of money.

On the other hand, a drug like acyclovir, which is used to treat (but not cure) herpes infections, costs about 12 cents per pill. Which is to say, basically nothing (in practice, this means a prescription can be filled for $4 at Walmart). It’s a perfect illustration of how generic medications can be profoundly cheap, and can help the lives of millions of people.

These two examples highlight a persistent tension in our healthcare system when it comes to medications. There’s a good reason for new drugs to be expensive: They are under patent, and the billions of dollars a company can make selling a drug while under patent is just reward — and inducement — for the hundreds of millions of dollars it takes to research and develop new drugs.

That said, not every new drug is necessarily better than the drugs that already exist for the same purpose. Generally speaking, the FDA requires new drugs to be tested based on their performance relative to a placebo, not in comparison to existing treatments. That kind of research, known as comparative effectiveness, has repeatedly shown that new, more expensive medications may not always be more effective at treating the condition than older therapies. (See for instance this study on statins, or this one on high blood pressure medications.)

Once a new drug is approved, the ticking patent clock means that pharmaceutical companies are highly incentivized to sell a new medicine to as many people as possible. That’s where the massive marketing campaigns — to the public and medical professionals alike — come in, all with the goal of creating a new blockbuster drug.

Introducing “Comparative Satisfaction”

Which brings us to our data. At Iodine, we’ve done a lot of research into the experience of people taking drugs, tracking the experience of more than 100,000 Americans across hundreds of drugs. Since it’s reported by ordinary people and not measured clinically, this data isn’t the same stuff that’s tracked in a randomized clinical trial — but it can tell us a lot about what it’s really like to take a medication in real life.

So we were curious about what could be called Comparative Satisfaction: measuring how satisfied people were with a handful of bestselling drugs with similar generic drugs that cost a fraction of the price.

We started with a list of bestselling drugs from April 2013 to March 2014, and chose four of the top 10 sellers: Abilify (for schizophrenia; 2013 sales: $6.9 billion), Nexium (for heartburn; 2013 sales: $6.3 billion), Crestor (for high cholesterol; 2013 sales: $5.5 billion), and Cymbalta (for depression; 2013 sales: $4.1 billion).

We then chose four generic drugs used to treat the same conditions: quetiapine (also sold under the brand name Seroquel), omeprazole (brand name Prilosec), atorvastatin (Lipitor), and venlafaxine (Effexor). Note: these aren’t a generic version of the same drug; these are generic versions of different drugs that treat the same conditions. They’re the sort of options that a doctor might consider when evaluating which medication to prescribe to a patient.

It turns out that the reported “worth it” scores for these five brand drugs and their generic similars is remarkably similar. The drug by drug comparisons follow a similar pattern: 75% of people taking Nexium say it’s worth it, compared to 80% of those taking omeprazole, while 60% say that Cymbalta is worth it, compared to 59% taking venlafaxine, and 44% rate Abilify as worth it compared to 45% saying that quetiapine is worth it. And when you aggregate these numbers across the more than 6,000 people who reported taking one of these drugs, we found that 61% say their under-patent, brand-name drug is worth it (and 22% say it’s not worth it) compared to 64% who say their off-patent, generic drug is worth it (and 20% say it’s not worth it).

In other words, when we measure the Comparative Satisfaction of the two sets of drugs, both separately and as a group, we found that they are nearly identical.

$23 Billion vs. $500 Million

So what to make of this finding? For one thing, it speaks of how hard it is to develop new drugs that perform better (in terms of human satisfaction) than existing treatments. But there’s a more profound implication when you consider that in 2013 about $23 billion was spent on the five patent-protected drugs. Some simple math shows that if those prescriptions were filled with the generic similars instead, the total cost would be less than $500 million — meaning a savings of about $22.5 billion to the US healthcare system.

Now there are plenty of caveats here. First, our data measures patient-reported satisfaction, not clinical efficacy, so we’re not able to say that the drugs are equally effective clinically. Second, it’s not clear that every person, or even most, who took the patent-protected drug would be a candidate to take the generic drug. For many there could be good reasons why the treatment protocol specified the more-expensive drug. Relatedly, we don’t know that the people taking the alternatives were actually taking the generic versions of them. Third, though our survey of 6,000 represents a large sample, it could be that a larger or different sample might find other results.

That said, we believe that Comparative Satisfaction can provide some compelling insights into the day-to-day reality of taking drugs, where real people often stop taking their medications not just because of side effects or perceived efficacy but also, oftentimes, because of the cost of the medication.

To paraphrase an oft-said line, if the most effective drug in the world is so expensive that people can’t afford to take it, it isn’t effective at all.