Potential drug-drug interactions in the intensive care Frequency, clinical relevance and improvement strategy

Drugs are an important part of medical treatment for patients admitted to the intensive care unit (ICU). During medication prescribing, errors can occur, resulting in adverse drug events (ADEs). In hospitalized patients, around 16% of ADEs are caused by a potential drug-drug interaction (pDDI).
ICU patients are particularly vulnerable to pDDIs due to polypharmacy and their compromised health. However, continuous monitoring facilitates effective and timely risk management of pDDIs.
Computerized decision support systems (CDSSs) can support clinicians to prescribe medication safely, by showing pDDI alerts. Yet, CDSSs can generate irrelevant alerts, leading to alert fatigue, alert overrides and reduced CDSS effectiveness. To address this, we hypothesized that tailoring pDDI alerts to the ICU setting could improve alert relevance and advance CDSS effectiveness, resulting in less exposure to clinically relevant pDDIs for ICU patients.
This thesis aims to answer the following questions:
1. What is the frequency of clinically relevant pDDIs in the ICU?
2. Which pDDIs are clinically relevant in the ICU setting?
3. Does tailoring pDDI alerts to the ICU setting reduce the frequency of administering clinically relevant pDDIs?
This thesis shows that 40% of the assessed pDDIs were considered not clinically relevant in the ICU. Still, ICU patients are frequently exposed to clinically relevant pDDIs, potentially resulting in patient harm. Therefore, the use of CDSSs to warn about these clinically relevant pDDIs is justified. Tailoring pDDI alerts to the ICU setting reduced exposure to clinically relevant pDDIs for ICU patients, improved patient monitoring, and decreased ICU length of stay.