Mental health randomised controlled trials (RCTs) base treatment benefit on a sole primary outcome, that is typically symptom-based in depression trials. However, research with stakeholders, including service-users and clinicians, often reports that although symptoms are important, but other domains such as quality-of-life, social, and occupational functioning are also prioritised. This evidence is not influencing current RCT practice, with sole primary outcomes remaining the norm. This short communication discusses how Multi-criteria Decision Modelling (MCDM) could offer a solution to this mismatch of expectations, by using stakeholder priorities for outcomes collected in RCTs and combining them with outcome data to produce stakeholder-weighted composites. These can be substituted for the primary outcome, to provide a complementary view of benefit from a broader, stakeholder-weighted multidomain lens. This could reveal nuances in treatment benefit and an understanding of treatment alignment with stakeholder expectations.

Unpredictable threats can sustain vigilance states that bias perception, cognition, and behaviour beyond immediate stimulus processing. Human screams, as evolutionarily salient alarm signals, are especially effective in eliciting defensive responses, yet their cortical dynamics and consequences for subsequent social regulation remain poorly understood. Using functional near-infrared spectroscopy (fNIRS), we dissociated anticipatory and stimulus-evoked cortical responses during exposure to either unpredictable human screams or laughter, and examined whether these effects generalize to interpersonal behaviour. Seventy-one participants completed two auditory blocks (Threat: screams; Safe: laughter) while prefrontal and temporal cortical activity was recorded, followed by a virtual interpersonal task measuring preferred distance and perceived discomfort. Unpredictable screams increased anxiety and elicited sustained anticipatory activation in the left temporal cortex, together with phasic responses in bilateral dorsolateral prefrontal and posterior temporal regions. In contrast, laughter induced more limited fronto-central activation and no anticipatory effects. Behaviourally, threat exposure was associated with earlier stopping of the approaching avatar, consistent with greater preferred interpersonal distance and defensive distancing. Discomfort during the interpersonal assessment was also elevated, particularly when the threat condition was presented first, suggesting a possible influence of session order or familiarization. Finally, exploratory analyses suggested that catastrophizing and perceived stress contributed to variability in neural and behavioural responses. Together, these findings indicate that unpredictable auditory threat induces sustained cortical vigilance that extends into subsequent social interaction, shifting interpersonal regulation toward defensive distancing.