94 PART THREE | DELIRIUM Table 2. Patients included in prediction algorithm. Delirium n = No delirium n = Total n = Study period 52 522 574 Extra delirium cases1 46 - 46 Total 98 522 620 1For adequate power in the development of the delirium prediction algorithm, data on the predisposing and precipitating factors of 46 patients who developed delirium between July 2012 and September 2013 were added to the original dataset. These data were only used for the development of this algorithm. Absolute risks at delirium reported in the article were calculated with the original dataset. Statistical analysis Statistical evaluation of differences between non-delirious and delirious patients was performedwitha χ2-test, theFisher exact test, or the students t-test,whenever appropriate. Because of themultiple comparisons an adjusted P-value of .01was considered statistically significant. To create a delirium risk prediction algorithm that can be easily implemented in the clinic we used a tree analysis method.22 All predisposing and the grouped precipitating factors for delirium were used in this tree analysis, irrespective of the χ2-test and students t-test results, topredict the riskof developingdeliriuminsubgroupsof patients. Thenumber of splits in the tree was chosen in order to minimize the cross-validated prediction error. Five-fold cross-validation was used for validation of the algorithm. For both the original and the cross-validated model the area-under-the-curve (AUC) was calculated. Data were collected in the web-based database system OpenClinica version 3.1.2. Statistical tests were performed with SPSS version 20.0. The prediction algorithm was constructed with the software package R program Rpart (version 3.1).
RkJQdWJsaXNoZXIy MjY0ODMw