Table 1 Proposed quality assessment of ML research for clinical practice.
From: Machine learning prediction in cardiovascular diseases: a meta-analysis
Algorithms |
Clarity of algorithms Propose new algorithms Select the proper algorithms Compare alternative algorithms |
Resources |
Reliable database/center Number of database/centers Number of samples (patients/images) Type and diversity of data |
Sufficient reported data |
Manuscript with sufficient supplementary information Letter or editor, short article, abstract Report baseline characteristics of patients |
Ground truth |
Comparison to expert clinicians Comparison to validated clinical risk models |
Outcome |
Assessment of outcome based on standard medical taxonomy External validation cohort |
Interpretation |
Report both discrimination and calibration metrics Report one or more of the following: sensitivity, specificity, positive, negative cases, balanced accuracy |
