In very sick patients (those generally filing for compassionate use) it's sometimes difficult to pin down whether the drug actually caused a problem or death. W

In very sick patients (those generally filing for compassionate use) it's sometimes difficult to pin down whether the drug actually caused a problem or death. Were the patients properly adherent? Did they have comorbidities? Were they simply too sick and died? A few deaths might cause the FDA to order additional trials which may actually not be necessary, to the tune of millions of dollars. If you were a company that spent years and hundreds of millions trying to bring a product to market, would you want to take that risk?