gms | German Medical Science

Introduction: Several single studies showed that computerized physician order entry systems (CPOE) may have the potential to reduce medication error and adverse drug events (ADE), and therefore, contribute to patient safety (e.g. [Ref. 1], [Ref. 2]). However, the design and evidence from these studies differ substantially and no systematic review with a formal meta-analysis has yet been performed on this topic.

Methods: We performed a systematic literature search in MEDLINE (1966-April 2006) and EMBASE (1976-April 2006). We included controlled field studies that evaluated the effect of CPOE systems on medication error, potential ADE and ADE, in which the CPOE system was used by physicians in direct care in order to order drugs. We included all types of CPOE systems independent of the level of decision support and all types of drugs and clinical settings. We performed a random effects meta-analysis to estimate the risk ratios with 95% confidence intervals (CI). We performed subgroup analyses for categorical factors such as clinical setting, patient group, type of drug, type of system, functionality, and study design. We investigated the publication bias.

Results: We identified 26 CPOE studies that fulfilled the inclusion criteria. 19 studies reported sufficient data in order to be included in the quantitative meta-analysis. Overall, CPOE systems reduced the risk for medication errors by 63% (CI: 53-72%, 19 studies), for potential ADE by 45% (CI: 11-66%, 8 studies), and for ADE by 56% (CI: 17-76%, 5 studies). A sub-group analysis did not show substantial differences. There was some evidence for a publication bias.

Discussion: Based on our meta-analysis, CPOE systems can reduce the risk for medication errors and ADE. However, studies differ substantially in their design, quality, and results. More randomised controlled trials from more sites are needed to improve the evidencebasis of health informatics [Ref. 3] and to identify the settings that CPOE systems are most useful in.