From: Guide to evaluating performance of prediction models for recurrent clinical events
Prediction bias (n) | Non-smoker | Current smoker | Ex-smoker |
---|---|---|---|
Negative binomial | |||
Female, previous exacerbations | 0.17 | 0.21 | 0.21 |
Male, previous exacerbations | 0.12 | 0.24 | 0.14 |
Female, no previous exacerbations | − 0.02 | − 0.03 | 0.02 |
Male, no previous exacerbations | 0.01 | 0.05 | 0.00 |
Zero-inflated negative binomial | |||
Female, previous exacerbations | 0.16 | 0.18 | 0.19 |
Male, previous exacerbations | 0.12 | 0.24 | 0.14 |
Female, no previous exacerbations | − 0.01 | − 0.01 | 0.03 |
Male, no previous exacerbations | − 0.01 | 0.04 | −0.01 |
Andersen-Gill | |||
Female, previous exacerbations | 1.07 | 1.34 | 1.25 |
Male, previous exacerbations | 0.72 | 0.96 | 0.91 |
Female, no previous exacerbations | 0.61 | 0.77 | 0.75 |
Male, no previous exacerbations | 0.43 | 0.57 | 0.54 |
PWP | |||
Female, previous exacerbations | 1.06 | 1.39 | 1.25 |
Male, previous exacerbations | 0.73 | 0.92 | 0.91 |
Female, no previous exacerbations | 0.61 | 0.76 | 0.73 |
Male, no previous exacerbations | 0.44 | 0.56 | 0.55 |