OBJECTIVES To examine, by a cross-sectional survey, how well prognostic prediction models for obstetric care were reported according to the TRIPOD checklist, and explore factors associated with poor reporting.

METHODS We searched PubMed for prognostic prediction model studies published on top general medicine or major specialty journals between January 2011 and February 2018. Teams of investigators independently identified eligible studies and used uniform TRIPOD data extraction checklist (22 items) and scoring rules to assess the reporting quality of included studies. The adherence proportion to each item was described, and multivariable logistic regression was applied to explore the association between methodological characteristics and poor reporting.

RESULTS Of 91 included studies, 74 (81%) studies were about prognostic prediction model development, 7 (8%) about external validation, 6 (7%) about incremental value, and 4 (4%) about model development and validation, with a median (25^th-75^th percentile) adherence of 40% (33-47%), 30% (30-47%), 47% (40-51%), 44% (40-49%), respectively. Overall, items reported less than 25% were title (5.5%), abstract (1.1%), study dates (16.5%), details of treatments (20.9%), blinding of outcome (24.2%) and predictors (2.2%), sample size (12.1%), predictor handling (7.7%), model-building procedures (2.2%), method for model performance (13.2%), flow of participants through the study (2.2%), how to use the model (11.9%), model performance (9.9%), implications (19.5%) and funding (5.5%); while better reported (>90%) were background and objectives, design and data source, participant eligibility, numbers for model development, and study limitations. The multivariable logistic regression found studies with larger sample size were inclined to be poorly reported (≥5000 vs. <5000: adjusted OR: 0.32, 95%CI: 0.11-0.96).

CONCLUSIONS The reporting of prognostic prediction models for obstetric care was poor according to the TRIPOD checklist, which might hinder transformation of available models in clinical practice. More efforts should be made to promote the implementation of the guideline and further improve its applicability for large-sample studies.