School-Based Education Programs for the Prevention of Child Sexual Abuse: A Cochrane Systematic Review and Meta-Analysis

Probably not a substantial claim: may not pass S3 (load-bearing).

The searches identified 21 prior reviews published up to August 2014.

“Our searches identified 21 reviews published up to August 2014.”

May combine two claims.

The review included 24 studies with a total of 5802 participants.

“Twenty-four studies with 5802 participants were included.”

• Also stated in: § Results of the Searches · plain

  “In total, this current review included 24 unique trials reported in 29 papers.”

• Also stated in: § Participants. · plain

  “A total of 5802 school-aged participants were included in the 24 trials.”

Probably not a substantial claim: may not pass S3 (load-bearing).

One previously included study (Pacifici, 2001) was excluded because it focused on sexual violence prevention in dating relationships for adolescents.

“The original included studies were reassessed and one (Pacifici, 2001) was excluded due to its focus on sexual violence prevention in the context of dating relationships for adolescents”

• cites Pacifici, 2001Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

New searches yielded 12,969 database records and 58 records from other sources.

“In the new searches, we identified a total of 12,969 records through database searching and a further 58 records from other sources”

Probably not a substantial claim: may not pass S3 (load-bearing).

After duplicates were removed, 10,218 records were screened by title and abstract.

“After duplicates were removed, we screened the titles and abstracts of 10,218 records”

Probably not a substantial claim: may not pass S3 (load-bearing). The quote only partly states this claim.

Three studies collected protective-behaviour data using a stranger simulation test.

“Of the 24 included studies, three studies reported collecting data on protective behaviours (i.e.demonstration of self-protective skills) (Fryer 1987a;Kraizer 1991;Poche 1988).”

• cites Fryer 1987aNot checked yet

  Not checked in the free run.

• cites Kraizer 1991Not checked yet

  Not checked in the free run.

• cites Poche 1988Not checked yet

  Not checked in the free run.

• AssumptionRests on an assumption: The stranger simulation test used to measure protective behaviours validly reflects children's self-protective skills in real-life abuse situations.

May combine two claims.

Six studies measured harm related to children's participation in school-based sexual abuse prevention programs, and three studies with 795 participants were included in the meta-analysis.

“A total of six studies measured harm in relation to children's participation in school-based sexual abuse prevention programs.We included three studies (795 participants) in the metaanalysis.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Seventeen trials used a wait-listed control group to compare program effectiveness.

“In 17 trials, the effectiveness of prevention programs was compared to that of a wait-listed control group.”

May combine two claims.

The standardized mean difference for vignette-based knowledge was 0.45 (95% CI 0.24 to 0.65), a moderate effect.

“The SMD was 0.45 (95% CI 0.24 to 0.65), indicating that those receiving treatment had an average 0.45 SD increase in applied knowledge as reflected in their responses to vignettes administered post-intervention, a gain of moderate effect size”

May combine two claims.

In the protective behaviours meta-analysis, heterogeneity was low to moderate and non-significant (I2 = 27%; Tau2 = 0.16; P = 0.24).

“In the analysis, heterogeneity approached the moderate range (I² = 27%; Tau² = 0.16) and was nonsignificant (P value = 0.24).”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Questionnaire-based measures were used in 21 of 24 included studies, and ten of these reported retention of knowledge from one to six months after intervention.

“Questionnaire-based measures were used in 21 of the 24 included studies.Ten of these studies reported on retention of knowledge over time, from one to six months post intervention.”

May combine two claims.

Earlier reviews had methodological weaknesses, including combining randomised and nonrandomised studies, pooling heterogeneous outcomes, and using weak statistical methods.

“These reviews were limited by methodological flaws such as pooling randomised and nonrandomised studies; aggregating diverse outcomes; and the use of suboptimal statistical methods.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Twenty-one studies reported questionnaire-based knowledge, using various measures.

“Of the 24 included studies, 21 reported questionnaire-based knowledge using a range of different measures.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Most studies (16 of 24) were conducted in North America, with the remainder in Europe, East Asia, and Central Asia.

“Most (16 of 24) were conducted in North America, the remainder in Europe, East Asia, and Central Asia.”

Probably not a substantial claim: may not pass S3 (load-bearing).

The 24 included studies comprised seven RCTs, 11 cluster-RCTs, and six quasi-RCTs.

“Of the 24 included studies, seven were randomised controlled trials (RCTs), 11 were cluster-RCTs, and six were quasi-RCTs.”

May combine two claims.

Complete data were unavailable for any study of retention of protective behaviours, so meta-analysis could not be performed.

“Complete data were not available for any of these studies.Meta-analysis could not be conducted.”

May combine two claims.

Narrative synthesis found that a small proportion of program participants experienced anxiety or fear, but these were mild rather than severe and short- rather than long-term.

“Narrative synthesis of included studies revealed that a small proportion of program participants experienced anxiety or fear but these (anxieties or fears) were mild rather than severe, and shortrather than long-term.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Fourteen interventions were brief, with a total duration of less than 90 minutes, and the remainder lasted from 90 to 180 minutes.

“Fourteen interventions were brief (i.e. less than 90 minutes total duration), and the remainder were longer, lasting from 90 to 180 minutes in total duration.”

May combine two claims.

Vignette-based measures were used in 12 of 24 studies; nine reported retention of knowledge, but none could be meta-analyzed because data were unusable or wait-list controls had already received the intervention.

“Vignette-based measures were used in 12 of the 24 included studies.Nine of these studies reported on retention of knowledge over time.None could be included in a meta-analysis because either the study did not provide data in a form useable in meta-analysis, or the study's wait-listed control group had already received the intervention.”

May combine two claims.

The review included 24 trials published in 29 reports, with 5802 child participants, approximately 98.8% from primary schools.

“This review reported on the effects of 24 trials (published in 29 reports) examining the effectiveness of school-based programs for the prevention of child sexual abuse including 5802 child participants of whom approximately 98.8% were from primary (elementary) schools.”

May combine two claims.

The protective-behaviour meta-analysis could include only Fryer 1987a and Poche 1988; Kraizer 1991 did not provide pass/fail results by group.

“Only the Fryer 1987a (n = 48; randomised controlled trial (RCT)) and Poche 1988 (n = 74;cluster-RCT) studies could be included in the meta-analysis for protective behaviours, as Kraizer 1991 (n = 670; cluster-RCT) did not report a breakdown of pass or fail scores for intervention and control groups.”

• cites Fryer 1987aNot checked yet

  Not checked in the free run.

• cites Poche 1988Not checked yet

  Not checked in the free run.

• cites Kraizer 1991Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

The earlier version of the review found 15 studies.

“The original review identified 15 studies (Zwi et al., 2007).”

• cites Zwi et al., 2007Not checked yet

  Not checked in the free run.

May combine two claims.

All included studies were conducted in school settings: 23 in primary schools and one in a special school for adolescents with intellectual disabilities.

“All studies were conducted in school settings: 23 in primary (elementary) schools and one in a special school for adolescents with intellectual disabilities (Lee & Tang, 1998).”

• cites Lee & Tang, 1998Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

Fifteen discrete prevention programs were identified across the trials.

“Fifteen discrete programs were identified.”

The quote only partly states this claim.

The current review found an SMD of 0.61 (95% CI 0.45 to 0.78; 18 studies, n = 4657) for questionnaire-based knowledge.

“The current review found a SMD of 0.61 (95% CI 0.45 to 0.78; 18 studies, n = 4657).”

May combine two claims.

At the time of the review, no study in this area had published a study protocol, and only one clinical trials register record was found for a school-based child sexual abuse prevention program.

“At the time of this review, no study in this area had yet published a study protocol, and we found only one clinical trials register record for a school-based education program for prevention of child sexual abuse”

Probably not a substantial claim: may not pass S3 (load-bearing).

In the original review, data extraction was done independently by two reviewers.

“Two review authors independently performed data extraction in the original review.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Between-study variability was examined with forest plots and the Tau2 and I2 statistics.

“We assessed heterogeneity using visual inspection of forest plots and two statistical measures (Tau² and the I² statistic), in order to describe and measure variability in the data”

Classified as probably not a substantial claim.

The review restricted inclusion to randomised, cluster-randomised, and quasi-randomised trials.

“We included studies if they were randomised controlled trials (RCTs), cluster-RCTs, or quasi-RCTs where participants were randomly allocated to the intervention or control group or were allocated using a quasi-random method such as day of the week, alphabetical order, or other sequential allocation such as class or school.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Two review authors independently screened titles and abstracts in the first two screening phases.

“In phase one, we imported titles and abstracts of articles identified in the searches into reference management software and two review authors independently screened them.”

• Also stated in: § Screening and selection of studies. · plain

  “In phase two, two review authors independently screened the titles, abstracts, and methodology sections of papers appearing to meet inclusion criteria.”

Probably not a substantial claim: may not pass S3 (load-bearing).

The meta-analysis used random-effects models to combine data.

“We used a random effects model to combine data.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Each study's risk of bias was evaluated with the seven-domain Cochrane revised tool.

“We assessed risk of bias in each study using the seven domains on the Cochrane revised 'Risk of bias' assessment tool”

May combine two claims.

The disclosure effect became less certain and was sensitive to different assumptions about clustering.

“making the result less certain, and indicating the effect of intervention programs on disclosure was sensitive to different assumptions regarding the effect of clustering on the results”

May combine two claims.

Knowledge scores may be affected by assessor bias or by contamination from group assessment, so better-controlled studies may produce smaller effects for this outcome.

“indicating that knowledge scores may be influenced by assessor bias or contamination from group assessment, or both, such that better controlled studies may generate lower effect sizes for this knowledge outcome”

May combine two claims.

Schools are suitable delivery settings for prevention programs because they allow universal implementation at relatively low cost without singling out higher-risk children.

“From a public health perspective (Rosenberg & Mercy, 1991) schools are appropriate systems for program delivery because programs can be implemented universally at comparatively little cost without stigmatising those at greater risk (Wurtele & Kenny, 2010)”

• cites Rosenberg & Mercy, 1991Not checked yet

  Not checked in the free run.

• cites Wurtele & Kenny, 2010Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Where ethnicity data were reported, participants were diverse, which increases the generalisability of the evidence and suggests widespread concern about child sexual abuse and its prevention.

“Where data were reported, participants were from a diverse range of ethnicities, increasing the generalisability of the evidence, and also suggesting that there is widespread concern about child sexual abuse and its prevention.”

Classified as probably not a substantial claim.

Classifying programs as active or passive may lead to erroneous conclusions about program effects.

“Classifying programs as either active or passive may therefore lead to erroneous conclusions about program effects”

May combine two claims.

The results suggest that intervention-group children showed greater factual knowledge about sexual abuse and prevention right after the program, with a moderate effect.

“These results suggest that children exposed to the interventions tend to display increased factual knowledge about sexual abuse and its prevention, when measured immediately after completion of the program, and the effect is of a moderate size”

• AssumptionRests on an assumption: The questionnaire and vignette measures used in the included studies validly measure knowledge of sexual abuse prevention.

May combine two claims.

Subgroup analyses for active or passive involvement were not conducted because programs could not be categorised that way; most were multifaceted.

“We did not conduct subgroup analyses for active or passive involvement as it was not possible to categorise programs in this way; most were multifaceted, involving combinations of approaches (see Table 2).”

May combine two claims.

Children's self-protective skills and factual and applied knowledge can be increased by participation in school-based sexual abuse prevention programs.

“Children's self-protective skills and factual and applied knowledge can be increased by participation in school-based sexual abuse prevention programs.”

Classified as probably not a substantial claim.

A public health approach to preventing child sexual abuse proposes multiple intervention types targeting offenders and potential offenders, children and adolescents, situations where abuse occurs, and communities.

“A public health approach for prevention of child sexual abuse proposes that multiple types of interventions are necessary, targeting: (i) offenders and potential offenders, (ii) children and adolescents, (iii) situations in which child sexual abuse is known to occur, and (iv) communities (Smallbone, Marshall, & Wortley, 2008).”

• cites Smallbone, Marshall, & Wortley, 2008Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

The review hand searched reference lists of previous reviews and included studies, and searched program evaluation databases and listservs for unpublished studies.

“We also hand searched reference lists of previous systematic and narrative reviews, and reference lists of included studies.We searched program evaluation databases such as the Promising Practices Network (RAND Corporation 2013), and Blueprints for Healthy Youth Development (Center for the Study and Running head: SCHOOL-BASED EDUCATION PROGRAMS FOR PREVENTION OF CHILD SEXUAL ABUSE Prevention of V”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Program content aligns with school health curricula, and schools provide a direct conduit to other prevention targets such as school personnel, parents, extended families, and communities.

“program content aligns with school health curricula (Sexuality Information and Education Council of the United States [SIECUS], 2004; Walsh, Berthelsen, Nicholson, Brandon, Stevens, & Rachele, 2013); and Running head: SCHOOL-BASED EDUCATION PROGRAMS FOR PREVENTION OF CHILD SEXUAL ABUSE schools provide a direct conduit to other prevention targets such as school personnel, parents, extended families”

• cites Sexuality Information and Education Council of the United States [SIECUS], 2004Not checked yet

  Not checked in the free run.

• cites Walsh, Berthelsen, Nicholson, Brandon, Stevens, & Rachele, 2013Not checked yet

  Not checked in the free run.

• cites Duane & Carr, 2002Not checked yet

  Not checked in the free run.

May combine two claims.

On balance, the programs do not appear to cause harm; they do not appear to increase or decrease children's fear or anxiety.

“On balance of evidence, programs do not appear to cause harm.That is, they do not appear to increase or decrease children's fear or anxiety.”

• Also stated in: § Conclusions · hedged

  “While we found few or no adverse outcomes in our review”

May combine two claims.

Using randomized or quasi-randomized designs does not guarantee that studies avoid substantial bias or are reported well enough for high-quality synthesis.

“While the use of randomized and quasi-randomized designs is laudable, their presence is no guarantee that the studies were carried out in ways that avoid substantial bias or that they are reported in ways that support high quality synthesis.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

The review searched 15 databases and updated searches to September 2014, with no restrictions on language or publication date.

“We searched the following 15 databases:”

• Also stated in: § Running head: SCHOOL-BASED EDUCATION PROGRAMS FOR PREVENTION OF CHILD SEXUAL ABUSE · plain

  “There were no restrictions on language or publication date.”

May combine two claims.

The translation of research into practice in this area is limited by lack of rigour in study methods and by failure to report the full range of child, intervention, and study design characteristics that could explain variations in program effects.

“On the whole, the translation of research into practice in this area is limited by lack of rigour in study methods and failure to report the full range of child, intervention, and study design characteristics that could possibly account for variations in program effects.”

May combine two claims.

Older children appeared to make greater knowledge gains than younger children on questionnaire-based measures, but not on vignette-based measures.

“Older children appeared to make greater knowledge gains than younger children when tested using questionnaire-based measures, but not with vignette-based measures”

Classified as probably not a substantial claim.

It is unknown whether skills tested in simulations involving approaches from strangers are the same as those needed for threats from familiar adults.

“First, it is not known if skills tested in the Running head: SCHOOL-BASED EDUCATION PROGRAMS FOR PREVENTION OF CHILD SEXUAL ABUSE context of approaches from strangers are the same as those required for children to deal with threats from familiar adults, who are the most common perpetrators of child sexual abuse.”

• AssumptionRests on an assumption: Familiar adults are the most common perpetrators of child sexual abuse.

May combine two claims.

School-based education programs fit within a public health approach by directly targeting potential victims and providing a conduit to other prevention targets such as teachers, parents, and extended families.

“School-based education programs fit within this approach, by directly targeting potential victims, and providing a conduit to other important prevention targets in communities such as teachers, parents, and extended families.”

May combine two claims.

Certainty about the disclosure result was limited because disclosures were poorly and inconsistently reported or not reported, and studies that reported disclosures did not use correct analytic techniques.

“However, we could not be certain of this result because children's disclosures were poorly and inconsistently reported, or not reported at all.Studies that did report disclosures did not use correct analytic techniques.”

May combine two claims.

Children exposed to the interventions showed increased factual and applied knowledge immediately after program completion, and scores were maintained for intervention participants one to six months after participation.

“Children exposed to the interventions displayed increased factual and applied knowledge about child sexual abuse and its prevention, when measured immediately after completion of the program, and scores were maintained for intervention participants one to six months after program participation.”

May combine two claims.

Results are strongest immediately after intervention, but there is more limited evidence that some effects are maintained over time.

“Results are best immediately postintervention, but there is more limited evidence that some effects are maintained over time.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Included studies were assessed on each risk-of-bias domain as low, high, or unclear risk, and the assessments were entered into RevMan and summarised in a risk-of-bias table.

“We assessed included studies on each domain as 'low risk', 'high risk', or 'unclear risk' with the latter indicating lack of information or uncertainty.We entered this information into RevMan and summarised it in a 'Risk of bias' table for each included study.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Cochrane Reviews have become more rigorous in identifying methodological limitations through risk-of-bias analyses, and CONSORT provides reporting guidance for randomized controlled trials.

“In the period since the original review was conducted (Zwi et al., 2007), Cochrane Reviews have become more rigorous in identifying methodological limitations in trials via risk of bias analyses, and the CONSORT statement has been developed to provide guidance on the reporting of randomised controlled trials (Shulz, Altman, & Moher, 2010).”

• cites Zwi et al., 2007Not checked yet

  Not checked in the free run.

• cites Shulz, Altman, & Moher, 2010Not checked yet

  Not checked in the free run.