Potential clinical relevant drug-drug interactions: comparison between different compendia, do we have a validated method?

Evaluation of Clinical Relevance of Drug-Drug Interaction Alerts Prior to Implementation

INTRODUCTION

Drug-drug interaction (DDI) alerts are often implemented in the hospital computerized provider order entry (CPOE) systems with limited evaluation. This increases the risk of prescribers experiencing too many irrelevant alerts, resulting in alert fatigue. In this study, we aimed to evaluate clinical relevance of alerts prior to implementation in CPOE using two common approaches: compendia and expert panel review.

METHODS

After generating a list of hypothetical DDI alerts, that is, alerts that would have been triggered if DDI alerts were operational in the CPOE, we calculated the agreement between multiple drug interaction compendia with regards to the severity of these alerts. A subset of DDI alerts (n = 13), with associated patient information, were presented to an expert panel to reach a consensus on whether each alert should be included in the CPOE.

RESULTS

There was poor agreement between compendia in their classifications of DDI severity (Krippendorff's α: 0.03; 95% confidence interval: -0.07 to 0.14). Only 10% of DDI alerts were classed as severe by all compendia. On the other hand, the panel reached consensus on 12 of the 13 alerts that were presented to them regarding whether they should be included in the CPOE.

CONCLUSION

Using an expert panel and allowing them to discuss their views openly likely resulted in high agreement on what alerts should be included in a CPOE system. Presenting alerts in the context of patient cases allowed panelists to identify the conditions under which alerts were clinically relevant. The poor agreement between compendia suggests that this methodology may not be ideal for the evaluation of DDI alerts. Performing preimplementation review of DDI alerts before they are enabled provides an opportunity to minimize the risk of alert fatigue before prescribers are exposed to false-positive alerts.

Predicting potential drug-drug interactions on topological and semantic similarity features using statistical learning

Drug-drug interaction (DDI) is a change in the effect of a drug when patient takes another drug. Characterizing DDIs is extremely important to avoid potential adverse drug reactions. We represent DDIs as a complex network in which nodes refer to drugs and links refer to their potential interactions. Recently, the problem of link prediction has attracted much consideration in scientific community. We represent the process of link prediction as a binary classification task on networks of potential DDIs. We use link prediction techniques for predicting unknown interactions between drugs in five arbitrary chosen large-scale DDI databases, namely DrugBank, KEGG, NDF-RT, SemMedDB, and Twosides. We estimated the performance of link prediction using a series of experiments on DDI networks. We performed link prediction using unsupervised and supervised approach including classification tree, k-nearest neighbors, support vector machine, random forest, and gradient boosting machine classifiers based on topological and semantic similarity features. Supervised approach clearly outperforms unsupervised approach. The Twosides network gained the best prediction performance regarding the area under the precision-recall curve (0.93 for both random forests and gradient boosting machine). The applied methodology can be used as a tool to help researchers to identify potential DDIs. The supervised link prediction approach proved to be promising for potential DDIs prediction and may facilitate the identification of potential DDIs in clinical research.