The deaths last spring of three patients in a clinical trial of an experimental cancer treatment made by Juno Therapeutics was tragic news and cause for reflection on FDA oversight of investigational new drugs. The company’s announcement yesterday of two additional deaths raises important questions about whether this oversight is being adequately exercised, supports growing calls for greater transparency around key events in the clinical drug development process, and underscores the need for rigorous clinical trials before drugs are approved.

When a pharmaceutical manufacturer or other sponsor is ready to begin testing an experimental therapy in humans, it must submit to the FDA an investigational new drug authorization, commonly called an IND. The clinical trial can begin as soon as 30 days after filing the IND. If the FDA becomes concerned about a major safety problem after its review of the IND — which is not nearly as exhaustive as its assessment of new drug applications prior to marketing approval — or such a problem emerges after the start of a trial, the FDA can issue what is a clinical hold. That means suspending the trial, at least until the hold is lifted. Clinical holds are quite rare. One review found that fewer than 30 had been issued between 2008 and 2014, despite thousands of drugs being tested under INDs at any time.

Read more: Two patient deaths halt trial of Juno’s new approach to treating cancer

In Juno’s case, the FDA’s hold lasted just three days. As my colleague Spencer Phillips Hey and I wrote in the BMJ, this period was atypically short. It was a missed opportunity to show that the agency had conducted a thorough review of the issues. We don’t know exactly what the FDA did or did not do because all of this regulatory activity happens outside public view. The existence of an IND, the details of a clinical hold, and the rationale for removing the hold are deemed trade secrets that the FDA is not allowed to routinely disclose. In fact, most of what we know about clinical holds are only filtered through press releases that manufacturers issue, which are unreliable sources of information.

The broad protection of trade secrets in this context comes at a steep price. Details on the review and resolution of clinical holds may provide lessons for other researchers seeking to improve the conduct of their trials, useful information for patients considering whether to enroll in a clinical trial (especially for of the drug at issue), and reassurance about the process that regulators are using to protect the public.

The Juno news also reinforces the importance of conducting rigorous clinical trials for new medicines. Prescription drugs can offer substantial benefits for some patients but can, of course, also be harmful. This is even true for drugs branded as “precision” medicines, or products like the one tested by Juno Therapeutics that the FDA has labeled as “breakthrough therapies.”

In recent years, numerous states have passed so-called “right-to-try” laws that encourage patients to seek access to experimental drugs outside of the clinical trial framework. In addition, libertarian activists and even some individuals associated with the incoming Trump administration continue to propose moving new medicines out into widespread use after only scant safety testing. That would increase the number of patients at risk for adverse outcomes, like the ones observed in the Juno trials, before we even know whether the drugs work.

Read more: Right-to-try laws could curtail the development of innovative new therapies

The best way to identify transformative new medicines, protect patients from unexpectedly dangerous drugs, and avoid wasting health care resources is by subjecting experimental products to well-designed clinical trials that enroll sufficient numbers of patients and test relevant clinical outcomes that can then be independently reviewed by the experts at the FDA. When severe, unanticipated problems arise, the FDA needs a transparent and systematic evaluation process that can provide public insight into what happened and why. That would contribute to the progress of science and the development of the next generation of safer, better therapies.

Aaron S. Kesselheim, MD, is an associate professor of medicine at Harvard Medical School and director of the Program On Regulation, Therapeutics, And Law (PORTAL) in the Division of Pharmacoepidemiology and Pharmacoeconomics at Brigham and Women’s Hospital.