Cancer patients in clinical trials should be told whether they’ve been given placebos or experimental drugs if their tumors reappear or get worse — and patients and researchers should be told if sponsors are worried experimental drugs may be triggering bad reactions, the FDA says in new draft guidance.

Using placebos in double-blind, randomized trials is a traditional way to protect against bias, but it also raises “both practical and ethical concerns,” the agency says. For one thing, giving a mere placebo to someone who is getting sicker could be perceived as tantamount to denying them care. For another, many experimental drugs may have toxic side effects and it can be dangerous to keep patients and/or researchers in the dark about whether a med or disease is to blame for symptoms.

“For example, in a blinded immunotherapy trial, a patient who develops adverse events … may receive unnecessary treatments,” the FDA says. “Maintaining the blind [status] after disease progression could also affect a patient’s subsequent therapy, potentially preventing a patient who had been on a placebo from receiving an approved therapy, or delaying or preventing the patient’s entry into other clinical trials.”

In an effort to assuage concerns, many recent cancer trials take an open-label approach where patients are given an already-approved treatment instead of placebos to compare against the experimental one. Other trials use an add-on approach—all patients are given standard treatment plus either a placebo or an experimental treatment.

But in some cases there aren’t other treatments available. In those rare instances, the FDA suggests trials “un-blind” patients if their disease recurs or worsens while taking placebos.

The draft guidance—released late last week—says both patients and investigators should also be “un-blinded” if an experimental treatment causes harmful side effects and the patient may need another med or even surgery. But it recommends patients be allowed to stay in trials even if they’re un-blinded.

The FDA notes that if sponsors insist on keeping trials “blind” throughout, at the least, they need to carefully explain this may happen—and the risks involved—in informed consent documents before patients sign up.

Read the draft guidance here.