The authors concluded that there was insufficient evidence to evaluate the diagnostic value of a screening physical examination in children without prior suspicion of child maltreatment. In view of the limited evidence found, the conclusions of this well-conducted review are likely to be reliable.

To evaluate the diagnostic value of a complete physical examination as a screening instrument to detect maltreatment in children without prior suspicion.

MEDLINE, EMBASE, PsycINFO, CINAHL, and ERIC were searched up to August 2013. Search strategies were reported. Lists of cited/citing references of relevant references were also consulted and field experts for contacted to identify additional studies. Studies were selected regardless of language or publication status.

Studies that evaluated a complete physical examination specifically performed as (part of) a screening procedure for child maltreatment and providing systematic documentation of the presence or absence of signs indicating child maltreatment were eligible for inclusion. Children from 0 to 18 years visiting any health care setting without prior suspicion of maltreatment were eligible. Complete physical examination was defined as minimally consisting of a visual inspection of the entire skin and oral cavity. Studies where physical examination was only performed to detect sexual child abuse were excluded. Examinations had to be conducted by a health care professional and had to be compared with one of the following reference standards: a court; the Child Protective Services; an expert panel; a forensic physician; or a self-report.

One study included a community sample of school-aged children in rural Egypt. The other two studies took place in an emergency department (in Italy and the USA). Ages of the children varied across the studies. All studies included mostly boys. Index tests and reference standard used varied across the studies. Studies were published from 1982 to 2003.

Studies were selected by two reviewers independently. Disagreements were resolved through consensus.

Quality assessment was performed by two reviewers independently using the QUADAS-2 tool. Disagreements were resolved via discussion or with a third reviewer.

Diagnostic accuracy data or data allowing the calculation of sensitivity and specificity, prevalence data and information on any reported harms were extracted by two reviewers independently. Disagreements were resolved via consensus or with a third reviewer.

Results were reported narratively and in tables.

Three studies were included in the review (476 to 10,175 individuals). All three studies had a high risk of bias related to reference standard. The reference standard was not independent of the results of the physical examination (incorporation bias) in two studies. Two studies did not clearly report the cumulative number of children with at least one positive finding based on the physical examination.

The prevalence of unconfirmed signs of maltreatment ranged from 7.8% to 14.6% of the children examined. The prevalence of signs of child maltreatment confirmed by a reference standard ranged from 0.8% to 13.5%. None of the studies reported adequate estimations of diagnostic accuracy. None of the studies reported any potential harm caused by a physical examination.

There was insufficient evidence to evaluate the diagnostic value of a screening physical examination in children without prior suspicion of child maltreatment.

The review addressed a clear question and was supported by clearly reported selection criteria. Multiple bibliographic sources were reported, and there were no restrictions on language or publication status. Attempts were made to reduce the risk of reviewer error and bias throughout the stages of the review.

Study quality was assessed with an appropriate tool and the results were reported. Given the limited evidence, the results were appropriately reported and synthesised, and were appropriately reflected in the conclusions.

The conclusions of this generally well-conducted review are likely to be reliable.

Practice: The authors stated that currently, clinicians should be aware that the diagnostic accuracy of a screening physical examination to detect child maltreatment was unclear and that child maltreatment could be both over- and under-detected. They stated that, although it was possible that a screening physical examination could identify emotional or sexual maltreatment in rare cases, the examination was aimed towards detecting physical abuse or neglect, and other forms of child maltreatment could be overlooked.

Research: The authors stated that a study should be conducted with a systematic physical examination and reference standard for a large, unselected group of children, at different levels of risk for maltreatment and in different settings. They suggested that the reference standard could be a thorough case review by an expert panel in combination with child, parent, informant or Child Protective Services report. They suggested that the reference standard could be performed in all children with a positive physical examination and in a random sample of children with a negative physical examination.

City of Amsterdam; Stichting Kinderpostzegels, The Netherlands.