Related Experiment Video
Updated: Nov 21, 2025
06:28
Biomechanical Changes Related to Low Back Pain: An Innovative Tool for Movement Pattern Assessment and Treatment Evaluation in Rehabilitation
Published on: December 13, 2024
891
A Bayesian Network Decision Support Tool for Low Back Pain Using a RAND Appropriateness Procedure: Proposal and
Adele Hill1, Christopher H Joyner2, Chloe Keith-Jopp1,3
1Sport and Exercise Medicine, Queen Mary University of London, London, United Kingdom.
JMIR Research Protocols
|January 15, 2021
Summary
This study outlines a method for building Bayesian networks to model expert clinical reasoning for low back pain (LBP). An internal pilot refined the process, highlighting software needs for effective clinical decision support tools.
Area of Science:
- Artificial Intelligence
- Clinical Decision Support
- Bayesian Networks
Background:
- Low back pain (LBP) presents a growing challenge, with increasing persistent pain and disability.
- Existing LBP decision support tools are limited by focusing on a subset of factors.
- Machine learning offers opportunities for developing advanced LBP management tools.
Purpose of the Study:
- Propose a modified RAND appropriateness procedure to elicit expert knowledge for constructing a Bayesian network for LBP.
- Report lessons learned from an internal pilot of the expert knowledge elicitation procedure.
Main Methods:
- Recruit expert clinicians from specialties relevant to LBP (orthopedics, rheumatology, sports medicine).
- Employ a four-stage process: online/face-to-face elicitation of variables, model structure, and probabilities, followed by validation.
- Utilize a modified RAND appropriateness procedure adapted for Bayesian network construction.
Main Results:
- Ethical approval obtained; internal pilot conducted with clinical colleagues.
- An alternating process of remote activities and in-person meetings proved necessary to avoid participant burden.
- Key lessons include the need for a bespoke online elicitation tool and clear operational definitions.
Conclusions:
- A method for constructing Bayesian networks representing expert clinical reasoning for musculoskeletal conditions has been proposed.
- An internal pilot demonstrated the method's potential success and identified software requirements for broader application.
- The refined process is suitable for developing clinical reasoning models for various conditions.
