Social cognition has not previously been assessed in treatment-naive patients with chronic schizophrenia, in patients over 60 years of age, or in patients with less than 5 years of schooling.

We revised a commonly used measure of social cognition, the Reading the Mind in the Eyes Test (RMET), by expanding the instructions, using both self-completion and interviewer-completion versions (for illiterate respondents), and classifying each test administration as ‘successfully completed’ or ‘incomplete’. The revised instrument (RMET-CV-R) was administered to 233 treatment-naive patients with chronic schizophrenia (UT), 154 treated controls with chronic schizophrenia (TC), and 259 healthy controls (HC) from rural communities in China.

In bivariate and multivariate analyses, successful completion rates and RMET-CV-R scores (percent correct judgments about emotion exhibited in 70 presented slides) were highest in HC, intermediate in TC, and lowest in UT (adjusted completion rates, 97.0, 72.4, and 49.9%, respectively; adjusted RMET-CV-R scores, 45.4, 38.5, and 34.6%, respectively; all p < 0.02). Stratified analyses by the method of administration (self-completed v. interviewer-completed) and by education and age (‘educated-younger’ v. ‘undereducated-older’) show the same relationship between groups (i.e. NC>TC>UT), though not all differences remain statistically significant.

We find poorer social cognition in treatment-naive than in treated patients with chronic schizophrenia. The discriminant validity of RMET-CV-R in undereducated, older patients demonstrates the feasibility of administering revised versions of RMET to patients who may otherwise be considered ineligible due to education or age by changing the method of test administration and carefully assessing respondents' ability to complete the task successfully.

Pharmacological treatment is widely used for post-traumatic stress disorder (PTSD) despite questions over its efficacy.

To determine the efficacy of all types of pharmacotherapy, as monotherapy, in reducing symptoms of PTSD, and to assess acceptability.

A systematic review and meta-analysis of randomised controlled trials was undertaken; 51 studies were included.

Selective serotonin reuptake inhibitors were found to be statistically superior to placebo in reduction of PTSD symptoms but the effect size was small (standardised mean difference −0.23, 95% CI −0.33 to −0.12). For individual pharmacological agents compared with placebo in two or more trials, we found small statistically significant evidence of efficacy for fluoxetine, paroxetine and venlafaxine.

Some drugs have a small positive impact on PTSD symptoms and are acceptable. Fluoxetine, paroxetine and venlafaxine may be considered as potential treatments for the disorder. For most drugs there is inadequate evidence regarding efficacy for PTSD, pointing to the need for more research in this area.