Method for evaluating prediction models that apply the results of randomized trials to individual patients

Table 2 Hypothetical data for some example patients using a treatment threshold (T) of 5%.

ID	Group	Event	Predicted risk if treated	Predicted risk if not treated	Risk reduction from treatment (D_i)	Treat?
1	Treatment	1	0.309	0.328	0.020	0
2	Control	0	0.352	0.396	0.045	0
3	Treatment	0	0.474	0.586	0.112	1
4	Control	1	0.219	0.350	0.132	1
5	Treatment	1	0.559	0.704	0.145	1
6	Control	1	0.360	0.409	0.049	0
7	Treatment	0	0.690	0.884	0.194	1
8	Control	1	0.406	0.477	0.071	1

The patients in bold are those whose treatment allocation is congruent with the recommendation of the prediction model at a T of 5%. Data from these patients would be used to calculate the outcome of decisions made by prediction modelling.

ISSN: 1745-6215