The authors concluded that parent-involved interventions for child survivors of sexual abuse had a small positive post-treatment effect compared with alternatives, such as child-only treatment, but this effect diminished over follow-up. Due to the small number of included studies and variation between the studies, caution may be required in interpreting these findings.

To evaluate parent-involved treatment of child sexual abuse victims.

The following databases were searched from 1980 to summer 2005: CINAHL, InfoTrac, MEDLINE, PsycINFO, Social Sciences Citation Index and Social Work Abstracts. Search terms were reported. The National Clearinghouse on Child Abuse and Neglect and the first four authors of relevant studies were contacted for information on additional studies. The reference lists of relevant reviews, (including a 2006 systematic review) were handsearched.

Comparative studies of parent-involved interventions for child survivors of sexual abuse were eligible for inclusion, provided the outcomes included child internalising symptoms (e.g. depression, anxiety) or externalising symptoms (e.g. behaviour problems), sexual behaviour problems or post-traumatic stress disorder (PTSD), and provided that effect sizes were calculable. Single group pre-test post-test studies were excluded.

The mean age of children in the included studies ranged from 4.68 to 11.4 years (where stated); and the proportion of female children ranged from 50% to 100%. The intervention (where described) consisted of eight to 20 sessions of cognitive-behavioural therapy delivered separately to parents and children, usually in an individual setting, plus (in some cases) joint parent and child sessions. All studies included comparison arms receiving active treatment (supportive therapy or treatment as usual); a small minority also included control arms receiving no treatment (waiting-list or community care). In some studies, control conditions included an alternative form of parental involvement.

Most studies used the Child Behavior Checklist to measure internalising and externalising symptoms (with or without other measures), and all studies that reported sexual behaviour used the Child Sexual Behavior Inventory. Results were based on parental report for most outcomes. Duration of follow-up (where reported) varied from three to 24 months. Most of the included studies were conducted by either of two research groups.

The authors did not state how the papers were selected for the review or how many reviewers performed the selection.

The following aspects of validity were assessed: randomisation method, blinding, attrition rate and use of intention-to-treat (ITT) analysis. Validity assessment was conducted independently by at least two reviewers, with discrepancies resolved by consensus.

All data were continuous. Post-test and follow-up mean differences between the treatment and comparison groups were calculated, with 95% confidence intervals (CIs). Where studies used multiple comparison groups, only the child-orientated treatment was included, and where outcomes were reported at multiple follow-up periods, only the first was used. Only comparisons between the intervention and active treatments (supportive therapy or treatment as usual) were included. For internalising and externalising symptoms, only Child Behavior Checklist scores were used.

Data were extracted independently by at least two reviewers, with discrepancies resolved by consensus.

Data were grouped by outcome and type of comparison. Studies were combined using Hedges' g formula to calculate standardised mean effect sizes (ESs) with 95% CIs. Both fixed-effect and random-effects models were used. Statistical heterogeneity was assessed using the χ² test. Moderator analysis was planned, to assess the impact on effect sizes of participant, intervention and methodological variables. Publication bias was assessed by calculating the number of studies that would need to be published to negate statistically significant results (fail-safe N), using the Rosenthal method.

Seven randomised controlled trials (RCTs) were included (516 participants). Three RCTs reported their randomisation method. About half the RCTs used blinded assessment, none blinded participants, most gave details of attrition rates and three reported using ITT analysis.

There was a statistically significant benefit for the intervention group at post-test in internalising symptoms (ES: 0.41, 95% CI: 0.21 to 0.61, six RCTs), externalising symptoms (ES: 0.32, 95% CI: 0.13 to 0.52, six RCTs), child sexual behaviours (ES: 0.31, 95% CI: 0.10 to 0.52, four RCTs), and PTSD outcomes (ES: 0.37, 95% CI: 0.14 to 0.55, five RCTs); and in child sexual behaviours at follow-up (ES: 0.45, 95% CI: 0.15 to 0.74, one RCT). These findings suggested small treatment effects.

No statistically significant difference between the groups was found at follow-up for internalising symptoms (two RCTs), externalising symptoms (three RCTs) or PTSD outcomes (three RCTs).

These are fixed-effect data, as no significant statistical heterogeneity was found for any outcome. The fail-safe N for statistically significant results ranged from four to 16. There were insufficient studies for moderator analysis.

Parent-involved interventions for child survivors of sexual abuse have a small positive post-treatment effect compared with alternatives such as child-only treatment. These effects diminish over follow-up.

The objectives and inclusion criteria of the review were clear and relevant sources were searched. It was unclear whether the search was limited by language or publication status. The methods used to select studies for inclusion were not reported. The processes of data extraction and validity assessment were conducted by more than one reviewer, but it was not stated whether they made decisions independently or how disagreements were resolved, so the potential for reviewer bias and error was unclear. Study validity was not explicitly used in the interpretation of findings. The choice of outcome measures was made post-hoc. These factors make it difficult to assess the reliability of the findings.

Suitable statistical techniques were used to combine studies and to assess for statistical heterogeneity and publication bias, but detailed outcomes data for the intervention and comparison groups would have aided interpretation of the results. Some potential limitations of the review were acknowledged, such as the small number of studies, variation between studies (particularly in control conditions) and the relatively low fail-safe Ns.

The review appears to have been well conducted in most respects, despite some omissions in reporting the methods. Due to the small number of included studies and variation between the studies, caution may be needed in interpreting the findings.

Practice: the authors stated that parent-involved treatment of child sexual abuse victims may need to be of extended duration or involve maintenance or booster sessions in order to maintain and increase benefits.

Research: the authors stated that more studies of parent-involved treatment of child sexual abuse victims were needed and that they should include follow-up assessment. Measures should be identified and studies conducted to assess the impact of these interventions on parents.

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.