Colpaert 2006.

Study characteristics
Methods	RCT ‐ individual. A prospective, controlled trial was conducted in two paper‐based units (PBUs; total of 14 beds (8 + 6)) versus one computerised unit (CU; 8 beds), 10 months after implementation of the intensive care information system (ICIS) in the latter unit. The objective of this study was to evaluate and compare the incidence and severity of medication prescribing errors (MPEs) between this CPOE unit and paper‐based units. Unit of allocation: patients Unit of analysis: prescriptions
Participants	22‐bed ICU of a tertiary university hospital, Centricity Critical Care Clinisoft (N = 90) IP adults (ICU)
Interventions	Intervention Technology Prescribing and order communication systems +  Intensive care information system (ICIS). Intervention: an intensive care information system (ICIS); that is, a computerised system specifically designed for the ICU that combines CPOE and a moderate level of CDSS. Control: paper‐based unit.
Outcomes	Prescribing errors Serious prescribing errors (potential to cause, or actually causing patient harm) The primary outcome measure was the difference in incidence and severity of medication prescribing errors (MPEs) in the CU versus the PBU. Secondary endpoints were univariate correlations between patient characteristics (APACHE II, renal failure, number of drug prescriptions (at screening day) and the number of MPEs.
Notes	No financial support stated. No trial number
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Patients were randomly assigned to either of these units by an independent nurse.
Allocation concealment (selection bias)	Unclear risk	Not described
Blinding of participants and personnel (performance bias) All outcomes	High risk	Not blinded
Blinding of outcome assessment (detection bias) All outcomes	Low risk	An independent panel, consisting of one clinical pharmacist, not involved in the registration part of the study, and two intensive care specialists, evaluated independently the severity of MPEs at least one month after screening. The panel was blinded for specific patient Grabar characteristics, as well as for patient group assignment.
Incomplete outcome data (attrition bias) All outcomes	Low risk	No loss to follow‐up was reported.
Selective reporting (reporting bias)	Low risk	The study protocol is not available but it is clear that the published reports include all expected outcomes, including those that were prespecified.
Conflict of interest	Low risk	Financial support not described
Other bias	High risk	Patients were randomly assigned to units, where there were two units for one arm and one unit for the other arm — i.e. patients were randomly assigned to study arms. Medical staff moved between the units (arms) on a one‐week basis. The outcome was measured as errors per prescription. There are possible clustering effects (e.g. of prescription within patient, and patient within unit). These possible effects are not accounted for by the analysis used.