GP155 Are clinical prediction rules in paediatrics validated? a literature review. (June 2019)

• Record Type: Journal Article
• Authors: Keegan Hickey, CA; Sarani, ZA; O'Neill, MB
• Is Part Of: Archives of disease in childhood. Volume 104:(2019)Supplement 3
• Page Start: A93
• Page End: A93
• Publication Date: 2019-06
• Subjects: Children -- Diseases -- Periodicals; Infants -- Diseases -- Periodicals; 618.920005
• DOI: 10.1136/archdischild-2019-epa.219
• Languages: English
• ISSNs: 0003-9888

Abstract

Abstract : Aims: To identify and appraise the geographical and temporal transportability of externally validated clinical prediction rules (CPRs) in the paediatric population. Methods: A Medline search was conducted from 2000–2015 for publications involving the derivation, internal or external validation of a CPR in children aged 0–21 years. Validation studies were included if they described a CPR which was developed within this time frame. Results: 193 studies were identified for inclusion. The selected studies describe 133 CPRs, across more than 10 paediatric specialities. 63 CPRs were not validated (47%). Of the 70 validated rules identified (53%), 25 were internally validated, 8 were cross validated and 53 were externally validated. Only 13 CPRs (10%) were both internally and externally validated and 21 CPRs (19%) were externally validated more than once. 15 internally validated studies used split sample methodology while only 5 performed bootstrapping regression methods. Of the 87 studies measuring external validity, the majority of studies (66%) were performed at a different site to the derivation. Only 5 studies were performed at the same site while 22 did not disclose this information. 43 externally validated studies included populations that were recruited in a different country to the derivation and 27 of those were on a different continent. Only two externally validated studies were shown not to have temporal transportability. Performance of the rules in validation was poorly reported, and if reported usually underperformed in comparison to the original. Conclusion: While many CPRs are developed, they are often not validated externally. Of those that have been validated, it is often difficult to uphold the same level of performance. The majority of externally validated CPRs assessed in this study meet geographical and temporal transportability which is a crucial feature in the CPR's ability to uphold performance amongst different populations.