Smeekens 2011.

Study characteristics
Methods	Study design: randomised controlled trial Unit of allocation: participant Unit of analysis: participant Adjustment for clustering: not applicable
Participants	Location: Utrecht, the Netherlands Setting: online module that participants could complete at the hospital (a university medical centre) or at home Sample size calculation: not reported Sample size: 38 emergency department nurses with permanent contracts; intervention group n = 19, control group n = 19 (Smeekens 2011, p 332) Mean age (SD): (i) intervention group = 41 years (SD = 9), (ii) control group = 41 years (SD = 11) (Smeekens 2011, p 332) Gender: (i) intervention group = 78.9% women, (ii) control group = 78.9% women (Smeekens 2011, p 332) Race/ethnicity: not reported Previous child protection training: not reported Years of experience: (i) intervention group = 9 years, (ii) control group = 9 years (Smeekens 2011, p 332) Previous experience with child maltreatment reporting: not reported Baseline equivalence: no significant differences in participant characteristics (Smeekens 2011, p 332)
Interventions	Name: Next Page (e‐learning module) Contents: 3 modules on child abuse: (i) recognition, (ii) acting, and (iii) communication Processes and teaching methods: online modules including: (i) simulations of clinical cases, (ii) video animations, and (iii) interactive elements Delivery mode: online Trainers and qualifications: module was hosted by a not‐for‐profit organisation called Augeo, whose website says interventions were designed in consultation with government, professional associations, and the International Society for Prevention of Child Abuse and Neglect (ISPCAN) Duration: 2 hours Intensity: participants complete the module in a minimum of 2 hours over a specified 2‐week window Intervention integrity: not applicable Comparison condition: no training
Outcomes	Eligible measures (outcome domain): performance in simulated cases (primary outcome: number of reported cases of child abuse and neglect, measured subjectively by participant responses to vignettes), comprising 8 cases based on real clinical cases with in vivo video‐recorded assessment and evaluated by an "expert panel of three paediatricians experienced in the recognition of child abuse" (Smeekens 2011, p 331) Ineligible measures (reason): self‐efficacy (self‐reported child abuse detection self‐efficacy; not prespecified in the protocol for this review), comprising 6 items corresponding to each of the 6 steps in the SPUTOVAMO‐R checklist, assessed on a 0‐to‐800‐millimetre visual analogue response scale Timing of outcome assessment: pre‐test (6 to 8 months before training), post‐test (2 weeks after training) (Smeekens 2011, p 331)
Notes	Funding: Augeo Foundation Author contact: no
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Comment: adequate description of the generation of the randomised sequence Quote: "participants were allocated to an intervention or a control group using a computer‐generated randomisation list created by an independent statistician" (Smeekens 2011, p 331)
Allocation concealment (selection bias)	Unclear risk	Comment: authors state that the trial was blinded, but it is unclear whether blinding refers to allocation concealment up until point of assignment. Study protocol cannot be used to verify this, as it has a different design to the trial reported in the journal article (see ClinicalTrials.gov NCT00844571). Quote: "design: blinded, randomised controlled trial" (Smeekens 2011, p 330)
Blinding of participants and personnel (performance bias) All outcomes	High risk	Comment: authors state that participants and personnel were not blinded, and would be able to identify if they were taking part in the intervention or not (self‐administered). Due to subjective nature of measures (pre‐post self‐report), possible performance bias without blinding. Quote: "owing to the nature of the trial it was not possible to blind the participants and the head researcher to randomisation" (Smeekens 2011, p 331)
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Comment: outcome assessors were blinded to participant group allocation, and assessed participant performance in simulations using a standardised assessment form Quote: "an expert panel of three paediatricians experienced in the recognition of child abuse, who were blinded to the allocation, evaluated the recorded performance ... . The case‐simulations were recorded on video and, after the completion of pre‐ and post‐test, the blinded expert panel scored the performance using a standardised assessment form which was designed to score quantity and quality of the questions posed by the nurse" (Smeekens 2011, pp 331–2)
Incomplete outcome data (attrition bias) All outcomes	High risk	Comment: 6/19 experimental and 7/19 control group participants were lost to follow‐up and not included in the analyses. Reasons for attrition reported (p 332): (a) "not scheduled to work in measurement period" (5 experimental; 6 control); (b) "unfinished e‐learning" (1 experimental); and (c) "participant left the role" (1 control). Authors performed additional analyses to examine the impact of attrition, but do not report data to support their statement that the results were unchanged. Quote: "to account for loss to follow‐up, both an intention‐to‐treat analysis with the pretest score carried forward and a multiple imputation analysis were performed. As the results were not essentially altered by these analyses we decided to present the analysis of the participants who performed the post‐test" (Smeekens 2011, p 332)
Selective reporting (reporting bias)	Low risk	Comment: study protocol is available, and outcomes align with those reported in the journal article. Self‐efficacy is assigned as a primary outcome in the study protocol, but is reported as a secondary outcome in the journal article.
Other bias	Unclear risk	Comment: timing of the post‐measure for the control and experimental groups may have differed, with the control group completing their post‐measure a short time (~2 weeks) prior to the experimental group. This poses a potential threat to internal validity. Also, additional potential sources of bias related to the specific study design have been identified.
Reliability of outcome measures (measurement bias)	Unclear risk	Comment: internal consistency for simulation outcome possibly not appropriate; however, authors report on the interrater reliability for the outcome assessors. No reliability data reported for the self‐efficacy measure (not eligible for this review). Quote: "the inter‐rater reliability for the three experts during post‐test was found to be 0.70 (95% CI 0.51 to 0.84, p value 0.000), which can be considered good" (Smeekens 2011, p 333)
Group comparability (selection bias)	Unclear risk	Comment: authors provide table of data comparing experimental and control participants on demographic and outcomes at baseline (no significant differences). However, the data are for the randomised participants (n = 38) and not the analysed participants (n = 25), which does not permit examination of whether analysed participants were comparable at baseline.
Contamination (contamination bias)	Unclear risk	Comment: unclear whether control and experimental participants worked in the same settings, thereby potentially leading to contamination

α: alpha; CPS: Child Protection Services; M: mean; NCATS: National Clinical Assessment and Treatment Service; NIH: US National Institutes of Health; Q&A: questions and answers; REDCap: research electronic data capture; SD: standard deviation; SPUTOVAMO‐R: a 9‐item Dutch checklist in which each letter in the acronym refers to 1 question in the checklist.