Citation and adherence to TRIPOD guidelines by published radiological prognostic models: systematic review

Area of Science:

• Radiology
• Medical Informatics
• Biostatistics

Background:

• Prognostic models are crucial in radiology for predicting patient outcomes.
• Standardized reporting guidelines, such as TRIPOD (Transparent Reporting of a multivariable prediction model for Individual Prognosis or Diagnosis), are essential for ensuring clarity and reproducibility.
• The extent to which these guidelines are adopted in high-impact radiological publications remains unclear.

Purpose of the Study:

• To assess the citation rates of the TRIPOD statement in prognostic models published in leading radiology journals.
• To evaluate the adherence of these published models to TRIPOD reporting recommendations.
• To identify specific areas of poor reporting within these prognostic models.

Main Methods:

• A systematic review of multivariable prognostic models developed using regression was conducted.
• Articles were identified from the top 3 indexed general radiological journals between December 2022 and May 2023.
• Reporting quality was assessed by checking for TRIPOD citation and scoring adherence to its domains.

Main Results:

• Only 4% of the 140 included articles cited TRIPOD, with only one submitting the checklist.
• Overall TRIPOD adherence was low, with a median score of 57%.
• Key areas with particularly poor reporting included title (2%), abstract (3%), and statistical analysis (5%). Only 38% of articles included a statistician author.

Conclusions:

• The majority of prognostic models in highly indexed radiological journals do not cite or adhere to TRIPOD recommendations.
• Improved adherence to TRIPOD is necessary for clear reporting of prognostic models.
• Radiological journals should mandate TRIPOD compliance for authors submitting prognostic models.