"At least do no harm" cannot be satisfied when failing to measure for possible harm.

Behavioral health professionals, in both their research and practice, have insufficiently prioritized understanding and communication of adverse experiences caused by psychological interventions for ethical and practical decision-making. This is due to a complex web of problems, including the lack of a foundational value commitment to "at least do not harm," confirmation bias, the lack of theoretical models of harm, the lack of valid measurement systems of potential adverse experiences, the lack of Popperian severe testing, and, relatedly, the lack of consistent use of adversarial collaborations and quality improvement systems. Proposals for future improvement are presented.