Sofia Priebe, 14, is slowly going blind. Her parents were devastated when they were told there’s no treatment for the genetic mutation that’s causing her retinas to deteriorate. For the dozen years since Sofia received that diagnosis, her mother has lived every parent’s nightmare—being powerless to help her suffering child.

Now a gene therapy for a similar form of blindness is expected to receive U.S. Food and Drug Administration approval this year, and Laura Manfre, Sofia’s mom, is holding out hope that her daughter may soon get treatment as well. “We don’t really care what it costs,” she says.

But how much is a miracle really worth? A million dollars? Five million? More? And who will pay and how? It’s one of the most vexing challenges confronting drug and insurance companies as modern medicine advances, spurred by research on the human genome. Spark Therapeutics Inc., which developed the gene therapy to cure a rare form of childhood blindness called RPE65-mediated inherited retinal disease, is among the first to face this question. Spark’s treatment, voretigene neparvovec, delivers a functioning piece of DNA directly to the eyes to preserve remaining sight and even restore some vision. Other companies, including GlaxoSmithKline Plc and BioMarin Pharmaceutical Inc., have also been grappling with the pricing problem.

Some new treatments, such as Spark’s retina drug, are intended to work with just one shot—promising a lifetime cure from a single, costly treatment. Insurers don’t dispute the worth of cures in the pipeline but say they’re not equipped to pay one large sum upfront. The U.S. health-care system is built around managing symptoms with prescriptions that insurers pay reimbursements for monthly: For example, medicines such as cholesterol-lowering Lipitor or acid-reflux drug Nexium are often taken over a period of many years and don’t deliver a permanent fix.

Insurers are “used to paying rent for health, and we’re asking them to buy a houseful of cure,” says Mark Trusheim, a visiting scientist at MIT’s Sloan School of Management who’s leading a working group to explore financing models for upcoming drugs, drawing from examples in the housing market and activist hedge funds.

Spark Chief Executive Officer Jeff Marrazzo sees his company’s pricing decision as precedent-setting, as it would be the first gene therapy approved in the U.S. Spark has spent about $400 million to create the treatment and now wants to be compensated for the efforts and huge risks it took during the research and development phase. How the payment debate plays out will determine not only whether patients will be able to gain access to these treatments but also how hard drugmakers will push to develop other transformative medicines. “Why is it that there are not more cures?” Marrazzo says. “It’s not because there are bad actors. It’s an industry full of people who react to incentive structures.” If compensation could be redesigned to reward one-time treatments over chronic treatments, “that’s where people would play,” he says.

One idea under consideration is to spread payments from insurance companies to drugmakers over years, like an annuity. Another is to have a money-back guarantee, so if a drug or treatment stops working for a patient, the manufacturer is on the hook to refund part of the cost to the insurer.

Payback contracts, or “value-based pricing” in pharma industry parlance, are already used in Europe. In the U.S. they’re trickier to execute, because there isn’t a single payer with which to negotiate and patients frequently switch insurance plans throughout their lives. Jean-Jacques Bienaime, CEO of BioMarin, proposes creating legislation requiring patients to carry the reimbursement obligation with them when they change jobs or insurers, to ensure drugmakers continue to be paid. BioMarin is working on a gene therapy for the blood disorder hemophilia, which Bienaime argues is easily worth millions of dollars per patient for a cure.

“The average cost of severe hemophilia A is about $500,000 a year, so $1 million upfront would cover two years—that’s not much, if they’re healed for life,” Bienaime says, adding that BioMarin hasn’t decided how to price its drug, which is going through trials.

MIT’s Trusheim is considering more radical payment plans, such as having the U.S. government buy an entire company instead of paying for its drugs. “The government could sell off the research and development arm of a company, like a Carl Icahn,” he says, referring to activist investors who take over corporations, often selling off units that they deem peripheral. The idea is to then have the feds provide the drug to Medicare and Medicaid patients and charge commercial insurers a cheaper price than they likely would have paid to a for-profit private company.

Other ideas that his group is investigating include government grants or prizes to developers of cures, or volume-purchase commitments such as those used by the Bill & Melinda Gates Foundation in developing nations, which guarantee manufacturers that a minimum amount of a drug will be bought if they successfully develop it. Trusheim’s group plans to publish its recommendations by early next year.

When designing payment models, the size of the patient population matters: The larger the potential impact, the more the system will strain to deal with the huge cost of one-time cures. That’s what Gilead Sciences Inc. learned in 2015, when it launched a cure for hepatitis C at $84,000 for a three-month regimen, or $1,000 a pill.

For anyone facing the prospect of liver cancer or a liver transplant, that $84,000 is a very good value, says Jim Meyers, Gilead’s executive vice president for global commercial operations. The price was in the ballpark of existing treatments, and “there wasn’t a payer we spoke to in market research that felt a price in the $80,000 to $85,000 range wasn’t acceptable,” he says.

The drug proved immensely popular—far beyond Gilead’s expectations—thanks to a quick endorsement by the American Association for the Study of Liver Diseases. About 3 million Americans have hepatitis C, and many wanted the cure immediately. “We started to see a flow of patients well above what any of us anticipated, and a price that made eminent sense to payers suddenly didn’t make sense,” Meyers says.

Rather than winning kudos for treating an intractable malady, Gilead quickly became a poster child for high drug prices, an image it’s fought to shed for the past two years. Meyers warns that the backlash has already spooked others. “What I hear is that they call it ‘the Gilead HCV experience,’ and by that they mean everything that it entails: the publicity, scrutiny, the payers first accepting then reacting, the Senate Finance Committee,” says Meyers. “I’ve heard, ‘Why bother?’ ”

Such market realities may already be reflected in new-drug prices. When GlaxoSmithKline brought its Strimvelis gene therapy to market in Europe last year for the “bubble boy disease” that leaves children without an effective immune system, it picked what’s arguably a bargain price: €594,000 ($634,000) for the one-time treatment, equivalent to two years of the enzyme replacement therapy that patients previously had to take for a lifetime.

Martin Andrews, Glaxo’s senior vice president for rare diseases, notes that the drug could have been valued higher, but he says the company had to bow to economic realities, including the big deficits in European governments’ health budgets. “We didn’t want to have the world’s most expensive therapy that nobody ever used,” he says.

Spark has a few more months to put a price tag on reversing progressive vision loss. Employees are researching court cases involving the value of sight. “How has our judicial system awarded for damages in the case of patients who have lost their vision?” asks Marrazzo. “That’s basically everyday Americans sitting on a jury and answering this question in a way that you can’t purely do with hard and fast economics.” Whether that will help its drug reach the market without consumer or government blowback remains to be seen.