Detecting adverse drug reactions on paediatric wards: intensified surveillance versus computerised screening of laboratory values.

Background: Adverse drug reactions (ADRs) contribute significantly to patient morbidity and mortality, as well as to costs for healthcare systems. Our aim was to evaluate the type and incidence of ADRs in a paediatric hospital population, comparatively ascertained by two different methodological approaches.
Methods: Our prospective study enrolled all patients admitted to two of the general children wards (46 beds) and the paediatric intensive care unit (6 beds) at the HELIOS Klinikum Wuppertal teaching hospital in Germany, over the study period of 3 months. We used two methods to detect ADRs. The intensified surveillance system relied on a trained physician conducting ward rounds and assessing patient charts. The computer-assisted screening of pathological laboratory parameters used values slightly below or above the age-specific normal range as a trigger signal for a potential ADR, which was subsequently assessed by trained personnel.
Results: By applying both methods simultaneously we observed that 14.1% of children experienced an ADR while they were hospitalised and 2.7% of children were admitted to hospital because of the ADR. Intensified surveillance resulted in the detection of 101 ADRs in 11.9% of patients, predominantly presenting with gastrointestinal symptoms, skin and CNS disorders; computer-assisted screening identified 45 ADRs in 5.7% of patients, mainly with drug-induced blood dyscrasia and liver damage. Furthermore, the ADRs detected by the intensified method were more severe, affected younger children and showed a closer causal attributability to the reaction than the ADRs observed by the computerised method. The spectra of drugs involved were similar, with the anti-infectives being suspected most frequently. The sensitivities of the intensified surveillance system and the computerised surveillance screening came to 67.2% and 44.8%, respectively, with computer-assisted screening having a specificity of 72.8%. The mean positive predictive value of the pathological laboratory values under surveillance by computer-assisted screening was 18.6%. Approximately 25% of ADR-related drugs administered were used for off-label indications.
Conclusion: Using the published literature for comparison, we found that ADRs occur as frequently in paediatric patients as in adult patients. Intensified surveillance and computerised surveillance applied in the paediatric setting show substantial differences in their detection specificities. A higher number of and more severe ADRs can be detected by intensified surveillance than by computerised surveillance, but require higher personnel resources.