Are computerised monitoring systems of value to improve pharmacovigilance in paediatric patients?

Abstract

Objective: The aim of the present study was to evaluate a computerised monitoring system (CMS) based on laboratory test results for the detection of adverse drug reactions (ADRs) on a paediatric ward.

Methods: A prospective, 6-month pharmacoepidemiological survey was performed on a 22-bed paediatric isolation ward. ADRs were identified by intensive chart review. In addition to spontaneous reporting by the treating physician, automatic laboratory signals generated by a CMS were evaluated for their association with ADRs. ADRs were classified by the affected target organs according to the WHO-ART system organ classes.

Results: A total of 73 ADRs were identified in 439 admissions (396 patients) by chart review. The CMS alerted 31 (42.4%) ADRs while 23 (31.5%) ADRs were found solely by treating physicians. Eight ADRs were detected by both approaches resulting in a total detection rate of 74% (compared with intensive pharmacovigilance). Out of a total of 27,434 laboratory tests performed routinely, 1,563 were classified as abnormal by the predefined CMS and used as the basis of alerts. The sensitivity of the system with respect to patients alerted was 90.3% and the specificity only 19.6%.

Conclusion: This study demonstrates that, using CMS, a different kind of mild adverse events were detected compared to the observation by the treating physician. The system presented appears to be sufficiently sensitive, but the specificity is too low to make it acceptable for physicians in daily practice. In children, clinically important ADRs can be detected best by intensified surveillance.