Related Experiment Video
Updated: May 12, 2026
An Experimental Paradigm for the Prediction of Post-Operative Pain PPOP
Published on: January 27, 2010
Prediction Models for Obstetric Anal Sphincter Injuries (OASIS): A Systematic Review and Critical Appraisal
Ellen S Mooren1, Jeroen van Bavel2, Anita C J Ravelli3
1Department of Obstetrics and Gynaecology, Hospital Rivierenland Tiel, Tiel, the Netherlands.
Background:
Obstetric anal sphincter injuries (OASIS) are complications with a risk of maternal morbidity. To estimate the individual risk of OASIS, prediction models have been developed.
Objectives:
Identifying studies on development and validation of prediction models for OASIS, with a critical assessment of methodology and clinical applicability.
Search Strategy:
This systematic review was performed following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines and the CHecklist for Critical Appraisal and Data Extraction for Systematic Reviews of Prediction Modelling Studies (CHARMS). PubMed, Embase, Cochrane Library and Scopus were searched up to 26 June 2024.
Selection Criteria:
Studies with description of model performance used for the development of prediction models.
Data Collection And Analysis:
Data extraction and assessment were performed by two independent researchers. The Prediction model Risk of Bias Assessment Tool (PROBAST) was used to assess risk of bias and applicability.
Main Results:
Of the 855 retrieved papers, 16 papers with 25 prediction models for OASIS were included. In these models, 2 to 15 variables were used. Model discrimination ranged from 0.64 to 0.83. All studies had shortcomings, particularly because of small or non-generalisable cohorts and were at risk of bias. Nine studies were of concern regarding clinical applicability. There were no geographical external validation studies of the 25 pre-existing models.
Conclusions:
The included prediction models for OASIS were of low or moderate quality and not applicable for use in clinical care yet. Future studies should focus on developing models based on larger generalisable multicentre cohorts, with clinically applicable predictors and with internal and external validation.
