The author stated that video feedback was not a moderator of treatment efficacy in social phobia and that individual treatments may be superior to group treatments, regardless of treatment type. In view of the limited search, poor reporting of review methods, lack of validity assessment and largely unexplored heterogeneity between the studies, the conclusions do not appear reliable.

To evaluate video feedback and treatment format as moderators of treatment efficacy in social phobia.

PsycINFO was searched. Search terms were reported. Eight (named) psychology journals were handsearched. The search was restricted to studies in English published from 2000 to 2006.

Studies of adults with generalised social phobia diagnosed by Diagnostic and Statistical Manual of Mental Disorders (DSM) IV criteria were eligible for inclusion, provided they used psychological therapy conducted face-to-face in a controlled clinical setting using either a group or individual format. Depression outcomes were required to be measured using one (or more) of five widely accepted and validated self-report scales listed in the review. Studies were required to report means and standard deviations (SDs) both before and after treatment.

Various psychological interventions were used in the included studies, in most cases types of cognitive and/or behavioural therapy. Data relating to pharmacological treatments and non-active controls were excluded. The variables of interest in the review were video feedback and delivery format (group or individual). The review outcome was whether these variables were moderators of psychological therapy with significant impact on treatment effect sizes. The review also explored the potential effects of treatment length and treatment format on attrition levels.

The author stated neither how the papers were selected for the review nor how many reviewers performed the selection.

The author did not state that he assessed validity.

Data were extracted on depression outcomes and categorised by the variables of interest. Mean changes from baseline within each category (treatment group) in each primary study were calculated and converted to effect sizes using a variant of Cohen’s d. Where multiple outcome measures were reported for a single treatment group, the mean d value for each group was calculated. The author did not state how many reviewers performed the data extraction.

The d values of individual treatment groups were pooled and weighted by sample size to calculate an overall weighted mean effect size. Q and I² statistics were used to assess within-group heterogeneity. The Q statistic was then used to assess whether there was statistically significant between-group heterogeneity, which would indicate that the variable of interest was a significant moderator of overall treatment effect.

Fourteen studies (18 study groups) were included (n=511, range 10 to 65). Attrition levels ranged from nil to 30 per cent. Most studies failed to use intention to treat analysis.

When video feedback groups (n=five) were compared with non-video feedback groups (n=13), weighted mean effect size was 0.15 higher in the former. Effect size was considered negligible. There was statistically significant within-group heterogeneity in both groups (video feedback groups Q=23.76, p<0.05, I²=83% and non-video feedback groups Q=60.54, p<0.05, I²=80%). There was no statistically significant difference in Q value between the groups (2.16, p>0.05). Video feedback did not significantly influence treatment effect.

When individual format groups (n=eight) were compared with group format groups (n=10), weighted mean effect size was 0.32 higher in the former. Effect size was considered small. There was statistically significant within-group heterogeneity in both groups (individual format groups Q=23.11, p<0.05, I²=70% and group format groups Q=51.95 p<0.05, I²=83%). There was a statistically significant difference in Q value between the groups (11.04 p<0.05). Treatment format significantly influenced treatment effect, favouring an individual format.

Post hoc analyses of attrition rates and treatment length were also reported in the review.

Video feedback was not a moderator of treatment efficacy in social phobia, but individual treatment may be superior to group treatment regardless of treatment type.

The objectives and inclusion criteria of the review were clear. The search was limited to one database and selected journals, and only published studies in English were eligible, which meant that some studies may have been missed and that the review was prone to language and publication biases. Methods used for study selection and data extraction were not described: as the review had a single author it appeared unlikely that these processes were conducted independently by more than one reviewer. It did not appear that study validity was assessed. These factors mean that the potential for bias and error in the review is unknown. No details were reported about important characteristics of the included studies (such as study designI. It was unclear whether the results of the pooled analysis were meaningful, given the high levels of unexplained inter-group heterogeneity and the uncontrolled nature of the comparisons. Potential sources of bias in the review were well addressed in the discussion section.

In view of the limited search, poor reporting of review methods, lack of validity assessment and unexplored heterogeneity between the studies, the author's conclusions do not appear reliable.

Practice: The author stated that both individual and group treatments should be considered a first-line intervention for social phobia and suggested that video feedback may be more effective if limited to individuals primarily with speech fears and distorted self-perception.

Research: The author stated that future studies should investigate the cause of study attrition.

Not stated.

This is a critical abstract of a systematic review that meets the criteria for inclusion on DARE. Each critical abstract contains a brief summary of the review methods, results and conclusions followed by a detailed critical assessment on the reliability of the review and the conclusions drawn.