Prediction models for skin tears in the elderly: A systematic review and meta-analysis
Siyue Fan,Hongzhan Jiang,Jiali Shen,Huihui Lin,Liping Yang,Doudou Yu,Mingqi Zhang,Nengtong Zheng,Lijuan Chen
DOI: https://doi.org/10.1016/j.gerinurse.2024.06.030
Abstract:Background: The prevalence of risk prediction models for skin tears in the elderly is growing; however, there is still debate regarding the usefulness and suitability of these models for clinical use and additional study. Objective: The purpose of this work is to perform a systematic review and meta-analysis of published research on skin tear risk prediction models in the elderly. Methods: We conducted a comprehensive search of various databases, including Cumulative Index to Nursing and Allied Health Literature (CINAHL), Embase, PubMed, Web of Science, MEDLINE, Scopus, The Cochrane Library, Wanfang Database, China Science and Technology Journal Database (VIP), and China National Knowledge Infrastructure (CNKI), from the beginning until November 27, 2023. Data extraction from the chosen studies encompassed various elements, such as study design, sample size, outcome definition, data source, predictors, model development, and performance. The assessment of bias and applicability was conducted using the Prediction Model Risk of Bias Assessment Tool (PROBAST) checklist. The Transparent Reporting of a Multivariable Prediction Model for Individual Prognosis or Diagnosis (TRIPOD) checklist was utilized to assess the transparency in reporting the prediction models-a meta-analysis of the most common predictors to assess predictor reliability. In addition, a narrative synthesis was carried out to provide an overview of the qualities, bias risk, and effectiveness of the current models. The reporting procedures of this meta-analysis conformed to the 2020 Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA 2020) statement. Results: Out of the initially retrieved 1499 studies, this review included eight prediction models from eight selected studies. All the studies employed logistic regression to develop prediction models for skin tears. The prevalence of skin tears in the elderly varied from 3.0% to 33.3%. Senile purpura and a history of previous skin tears were the most commonly utilized predictors. The reported values for the area under the curve (AUC) ranged from 0.765 to 0.854. All the studies exhibited a high risk of bias, primarily due to inadequate reporting in the outcome and analysis domains. Furthermore, serious questions concerning their applicability were highlighted by four studies. Conclusion: Based on the PROBAST checklist, the current models for predicting skin tears in the elderly showed a high risk of bias. The development of new prediction models with bigger sample sizes, appropriate study designs, and external validation from multiple sources ought to be the primary focus of future research. Patient or public contribution: There was no patient or public contribution to this systematic review. Registration: PROSPERO registration number: CRD42023494387.