voices, there is a need to tighten conceptualization and measurement of outcome. Psychological therapy trials for psychosis have been criticized for using outcome measures of symptom intensity used in antipsychotic drug trials, not well suited to the aims of psychological therapies of reducing impacts of symptoms on emotional well-being and functioning rather than their occurrence.96 The main voice-specific outcome measure used has been the Psychotic Symptom Rating Scales (PSYRATS)97 (see Woodward et al, this issue). Although capturing a number of dimensions, this measure has limitations in giving a total score which sums a series of only modestly intercorrelated scales, and which introduces noise from variables not targeted by therapy (such as voice frequency, location, and extent of negative voice content), while individual items are comprised of single 5-point ratings, likely to lack sensitivity to change when used instead. Indeed, different methods of scoring have been used between trials (single item, factor, or total score), reducing comparability of findings. Better measures are needed to capture the impact of voices on emotional well-being and functioning. Multiple item measures of the subjective impact of psychosis have been developed in recent years, which are likely to be more sensitive, but have yet to be reported on in trials, and are not voice specific. The application of these measures, and development of a voice-specific measure, are required to provide a more accurate estimate of targeted voice outcomes posttreatment and at follow-up.