Liu Yi, Jing Qing, Bian Yuan, Yan Junfeng
ObjectiveTo evaluate the effectiveness of self-established 33 triggers in detecting drug adverse events (ADE) in children and explore child-individual factors possibly associated with ADE.MethodsTwo hundred medical records of inpatient in Sichuan Provincial People′s Hospital, who discharged from January 1, 2017 to September 30, 2017, were sampled using the sampling method recommended by the global trigger tool white paper of U.S. Institute for Healthcare Improvement. The medical records were detected using the self-established 33 ADE triggers. If there were positive triggers, whether the children developed ADE, the classification, grading, and causal judgement of ADE, types of common drugs causing ADE in children, and the association between the individual factors in children and ADE were further analyzed.ResultsIn the 200 medical records, 200 children were involved, including 128 males and 72 females, aged from 38 minutes after birth to 18 years with the average age of 6 years. The time of hospital stay was 2 to 43 days with the average time of 10 days. Of the 200 medical records, 128 had at least 1 positive trigger, and the positive trigger rate was 64.0% (95%CI: 57%-71%). Of the 33 triggers, 29 were triggered a total of 394 times, and 98 ADE were detected, involving 41 children. Thus the detectable rate of the triggers was 20.5%(41/200), and the positive predictive value (PPV) of the triggers in detecting ADE was 24.9% (95%CI: 20.1%-28.0%). Ninety eight ADE could be classified into 12 categories, and ADE such as abnormal white blood cell count (21.4%, 21 times), skin lesions (11.2%, 11 times), and abnormal platelet count (10.2%, 10 times) were more common. The proportion of ADE that were grading as grade 1, grade 2, and grade 3 were 39.8% (39 times), 56.1% (55 times), and 4.1% (4 times), respectively. In the causality determination, 4 (4.1%), 73 (74.5%), 19 (19.4%), 1(1.0%), and 1 (1.0%) ADE were determined as certain, probable, possible, conditional, and unassessable. Ninety eight ADE involved 18 classes (including 52 kinds) of drugs, and the frequency of drug occurrence ranged from 1 to 16 times, with a total of 143 times. Anti-tumor drugs (44.1%，63 times) appeared most frequently, and followed by antibacterial agents (23.8%，34 times). Logistic regression analysis showed that hospital stay and diagnosis of leukemia were positively correlated with the likelihood of ADE occurrence.ConclusionsThe self-established 33 triggers could effectively monitor ADE in children, despite that further improvements were necessary. The individual factors in children that might be associated with ADE were more hospitalization days and suffering from leukemia.
