Table 1.
Predictor variables identified from the univariate analysis stage, and from other studies, only 8 of which were taken forward to the multivariate analysis stage (indicated with *).
Figure 1.
Recruitment diagram for the Brighton and Validation datasets.
Table 2.
Patient characteristics and co-morbidities in the Brighton and validation (European) data sets.
Table 3.
Results of multivariate analysis.
Figure 2.
ADR rate according to ADR risk score.
The BADRI risk model was applied to all 690 patients from the Brighton dataset (A), and 483 patients from the European dataset (B). The ADR rate is calculated as the proportion of patients in each scoring category that suffered an ADR. For both datasets there is a general increase in the ADR rate as the risk score increases.
Table 4.
Accuracy of the BADRI risk model as applied to the Brighton dataset using various cut-off values (risk scores).
Table 5.
Accuracy of the BADRI risk model as applied to the European dataset using various cut-off values (risk scores).