Development and validation of PRE-DELIRIC (PREdiction of DELIRium in ICu patients) delirium prediction model for intensive care patients: observational multicentre study

Dear Dr. Prabhakaran,

Many thanks for your interest in our article describing the development and validation of a delirium prediction model for ICU patients. Variables which were recognized as potential risk factors for the development of delirium in ICU patients were collected. Since lorazepam was recognized as an independent risk factor we included this, as well as the overall group of sedatives. However, lorazepam is not often used in our hospital (and in the Netherlands) mostly because of its long half-life. Therefore our prevalence rate was very low. After exclusion lorazepam as an individual predictor we included lorazepam in the group of sedatives, as noted in Appendix A. So the variable lorazepam is not completely lost as predictor. It was not our purpose to restrict the group of sedatives to midazolam, propofol or lorazepam but these are the most commonly used drugs for sedation in the Netherlands.

Concerning the second point, indeed we excluded the variables ‘alcohol abuse’ and ‘dementia’ because of the low prevalence in our population. Since the delirium incidence in this group (Appendix B) is very high, we feel that a delirium prediction model for these patients would be redundant. If we would include both variables in the modeling process this would result in dilution of the effects of the other co-variates. We advise to take preventive measures in patients with a history of alcohol abuse or dementia, irrespectively their predicted delirium chance.

Mark van den Boogaard and Peter Pickkers