Prediction Tools in Spine Surgery: A Narrative Review

Area of Science:

• Spine Surgery
• Medical Informatics
• Clinical Prediction Models

Background:

• There is a rising number of prediction models being developed for spine surgery.
• Prognostic tools and risk calculators are increasingly recognized for their role in shared decision-making between patients and clinicians.
• The development of these models utilizes various methodologies.

Purpose of the Study:

• To provide a comprehensive overview of the existing literature on the development and validation of prediction models specifically within the field of spine surgery.
• To highlight the importance of external validation for ensuring the generalizability and appropriate application of these models.
• To synthesize current knowledge on prediction model methodologies in spine surgery.

Main Methods:

• This study conducted a narrative review of published literature.
• The review focused on articles describing the development and external validation of prediction models in spine surgery.
• Literature search terms encompassed prediction models, risk calculators, prognostic tools, and spine surgery.

Main Results:

• The literature shows a significant increase in the development of prediction models for spine surgery.
• External validation is consistently emphasized as a critical step for assessing model performance in new populations.
• A variety of statistical and machine learning methods have been employed in model development.

Conclusions:

• Prediction models are becoming more prevalent in spine surgery, offering valuable insights for clinical practice.
• Rigorous external validation is essential for the trustworthy implementation of these prognostic tools.
• Further research should focus on standardizing validation processes and ensuring the clinical utility of developed models.