Ethical researchers will prematurely stop and publish a "no effect" result for any medicine that has bad effects. That is, unless the effects are so bad that they can't stop fast enough, and they get to measure a statistically relevant harm.
Thus seeing "no effect" places an upper bound on how good an effect you will see as "none", and a lower bound as "won't kill most people quickly". That's not a good space to gamble into.
It took almost an year until the scientific community settled down that HQC for treating COVID-19 is harmful. The actual formal result is still "no effect and doing more studies is anti-ethical" because nobody can tell exactly how harmful it is. Probably nobody will ever be able to tell (and that's a good thing).