TRIPOD+AI: an updated reporting guideline for clinical prediction models
New update promotes best practice in this important area of clinical research Clinical prediction models emerged in the 1990s as tools to support medical decision making through individual diagnostic and prognostic predictions based on structured clinical information. Clinical prediction rules such as the FeverPAIN score for pharyngitis1 or the PECARN rule for children with head trauma2 are based on prediction models and aid clinicians in prescribing antibiotics and ordering computed tomography (CT) scans, respectively. In a linked paper (doi: 10.1136/bmj‑2023‑07837), Collins and colleagues introduce TRIPOD+AI, an updated version of the TRIPOD statement to improve the reporting of studies on the development and evaluation of clinical prediction models.3 Transparent, accurate, and complete reporting is a prerequisite to any form of quality assessment of a study—including evaluating the risk of bias and applicability of study results—and increases the value and usability of scientific re