When the FDA is deciding whether to approve new drugs, one of their ethical responsibilities is clear—to deny treatments that involve more risk for patients than expected benefits. Of course, there are other reasons beyond mere morality for the agency to uphold this responsibility. For one, risky drugs cause undue harm to patients, potentially leading to lawsuits, negative publicity, and fatalities. As such, the FDA has stipulated a standard threshold of 2.5%—representing the maximal chance that drug trials might be false positives, essentially meaning that they are willing to consider drugs that are shown to have a less than 2.5% of being defective—for approving drugs intended to treat serious conditions and rare disease. While this risk threshold is essential for keeping patients safe from risky and defective drugs, could a risk threshold so low prevent effective drugs from reaching the market, thereby harming patients?

Could a risk threshold so low prevent effective drugs from reaching the market, thereby harming patients?

Two MIT researches seem to think so. In their recently published paper on the FDA’s drug approval process, economist Andrew Lo and grad student Vahid Montazerhodjat contend that the agency’s considerations are weighted more toward avoiding Type I error (approving ineffective therapies) than Type II error (rejecting effective therapies).1 As such, the FDA’s current policy deprives direly ill patients of treatments that could improve both their immediate condition and long-term prospects for survival.

Take pancreatic cancer, for example. Patients diagnosed with this devastating malady experience 5-year survival rates of barely 7%. The 5-year survival rate for prostate cancer, meanwhile, is 98.9%.2 Yet new drugs for both diseases—one quite treatable, the other nearly amounting to a death sentence—must pass the same stringent risk barrier for approval. Such a policy prevents patients suffering from numerous devastating conditions from accessing drugs whose benefits far outweigh their risks.

Lo and Montazerhodjat propose a simple remedy: adjust the risk threshold to a level proportionate to the severity of the disease. Employing Bayesian Decision Analysis to calculate the ideal trade-off between Type I and Type II
errors, they calculated that any pancreatic cancer treatment with less than a 27.9% of being defective should receive approval.3 Raising the acceptable risk threshold by a factor of more than ten would greatly increase the number of drugs on the market whose benefits outweigh potential harm for pancreatic cancer patients.4  Patients suffering from other severe conditions—such as lung cancer, liver cancer or cirrhosis of the liver—would also stand much to gain from a risk-adjusted approval policy.

The overall thrust of the paper is aimed at holding the FDA to one of its central ethical responsibilities—to approve treatments whose expected benefits outweigh the potential risk for patients. Though the FDA’s responsibility to minimize Type II error is less obvious than its need to minimize type I error due to the more visible consequences of the latter (e.g. patients dying from faulty and unsafe medication), the moral burden on the agency is equivalent in both cases. As such, the FDA should modify their drug approval process along the guiding ethical principle that any drug whose potential benefits outweigh possible risks should be approved. Under such a new process, the risk calculus that goes into approving new drugs must account for the relative severity of the condition the drug is intended to treat. Therefore the current “one size fits all” risk threshold for drug approval is both unethical and undesirable.

  1. Lo, Andrew and Vahid Montazerhodjat. “Is the FDA too conservative or Too Aggressive?” The National Bureau of Economic Research (August, 2015).
  2. “Fact Sheets: Pancreas Cancer.” And “Fact Sheets: Prostate Cancer.” National Cancer Institute. http://seer.cancer.gov/statfacts/html/pancreas.html and http://seer.cancer.gov/statfacts/html/prost.html. Accessed October 10, 2015.
  3. Lo, Andrew and Vahid Montazerhodjat. “Is the FDA too conservative or Too Aggressive?” The National Bureau of Economic Research (August, 2015).
  4. Lo, Andrew and Vahid Montazerhodjat. “Is the FDA too conservative or Too Aggressive?” The National Bureau of Economic Research (August, 2015).