Schouten Maartje C M, van Stel Henk F, Verheij Theo J M, Houben Michiel L, Russel Ingrid M B, Nieuwenhuis Edward E S, van de Putte Elise M
Division of Pediatrics, Wilhelmina Children's Hospital, University Medical Center Utrecht, Utrecht, the Netherlands.
Julius Center for Health Sciences and Primary Care, University Medical Center Utrecht, Utrecht, the Netherlands.
PLoS One. 2017 Jan 3;12(1):e0165641. doi: 10.1371/journal.pone.0165641. eCollection 2017.
To assess the diagnostic value of the screening instrument SPUTOVAMO-R2 (checklist, 5 questions) for child abuse at Out-of-hours Primary Care locations (OPC), by comparing the test outcome with information from Child Protection Services (CPS). Secondary, to determine whether reducing the length of the checklist compromises diagnostic value.
All children (<18 years) attending one of the participating OPCs in the region of Utrecht, the Netherlands, in a year time, were included. The checklist is an obligatory field in the electronic patient file. CPS provided data on all checklist positives and a sample of 5500 checklist negatives (dataset). The checklist outcome was compared with a report to CPS in 10 months follow up after the OPC visit.
The checklist was filled in for 50671 children; 108 (0.2%) checklists were positive. Within the dataset, 61 children were reported to CPS, with emotional neglect as the most frequent type of abuse (32.8%). The positive predictive value (PPV) of the checklist for child abuse was 8.3 (95% CI 3.9-15.2). The negative predictive value (NPV) was 99.1 (98.8-99.3), with 52 false negatives. When the length of the checklist was reduced to two questions closely related to the medical process (SPUTOVAMO-R3), the PPV was 9.1 (3.7-17.8) and the NPV 99.1 (98.7-99.3). These two questions are on the injury in relation to the history, and the interaction between child and parents.
The checklist SPUTOVAMO-R2 has a low detection rate of child abuse within the OPC setting, and a high false positive rate. Therefore, we recommend to use the shortened checklist only as a tool to increase the awareness of child abuse and not as a diagnostic instrument.
通过将筛查工具SPUTOVAMO - R2（包含5个问题的清单）的测试结果与儿童保护服务机构（CPS）提供的信息进行比较，评估其在非工作时间初级保健场所（OPC）对虐待儿童情况的诊断价值。其次，确定缩短清单长度是否会影响诊断价值。
纳入荷兰乌得勒支地区参与研究的OPC在一年内接诊的所有18岁以下儿童。该清单是电子病历中的必填项。CPS提供了所有清单阳性病例的数据以及5500例清单阴性病例的样本（数据集）。在OPC就诊后10个月的随访中，将清单结果与提交给CPS的报告进行比较。
共为50671名儿童填写了清单；108份（0.2%）清单呈阳性。在数据集中，有61名儿童被报告给CPS，其中情感忽视是最常见的虐待类型（32.8%）。该清单对虐待儿童的阳性预测值（PPV）为8.3（95%可信区间3.9 - 15.2）。阴性预测值（NPV）为99.1（98.8 - 99.3），有52例假阴性。当清单长度缩短为与医疗过程密切相关的两个问题（SPUTOVAMO - R3）时，PPV为9.1（3.7 - 17.8），NPV为99.1（98.7 - 99.3）。这两个问题分别是与病史相关的损伤情况以及儿童与父母之间的互动。
在OPC环境中，清单SPUTOVAMO - R2对虐待儿童的检出率较低，假阳性率较高。因此，我们建议仅将缩短后的清单用作提高对虐待儿童情况认识的工具，而非诊断工具。
