| Inclusion criteria An article must satisfy all of the following criteria to be included in the review. |
---|---|
1 | Compare at least one machine learning method and at least one statistical method (according to our definitions). Any number of hybrid methods can be compared but a machine learning method and a statistical method must be included. |
2 | Methods included should be prognostic (risk prediction) models for one, specific outcome in a medical/healthcare context. |
3 | Methods included must be used to predict survival outcomes. |
4 | The simulation study must have been used to compare the methods with a time-to-event outcome with censoring. |
5 | Methods must be evaluated and compared in terms of prognostic ability. |
6 | Methods must not be for modelling treatment effects, feature selection or genetic variant identification. |