A trial of automated decision support alerts for contraindicated medications using computerized physician order entry
First Article Review
A Trial of Alerts designed to reduce Inpatient Administration of Medications contraindicated due to Renal Insufficiency 
Secondary to the high use of multiple medications hospitalized patients are at risk for adverse drug events (ADEs). ADEs result in increase in cost of care as well as increased morbidity and mortality. Patients with renal insufficiency are at increased risk of ADEs from medication administration. The University of Illinois Hospital and Medical Center trialed a real-time CDS (clinical decision support) alert to reduce the incidence of medication administration contraindicated due to renal insufficiency. After a minimum safe creatinine clearance was established for each inpatient formulary medication, a “pop-up” alert recommending cancellation was developed for a medication order initiated for a patient whose estimated creatinine clearance was less than the safe creatinine clearance for the medication.
To further analyze the efficacy of the alert, the study reviewed the impact of patient demographics, degree of renal dysfunction, and the duration of housestaff training. After 14 months of the study, the 323 alerts were generated and the likelihood of receiving at least one dose of a contraindicated medication was reduced from 89% to 47% (p<0.0001). This 42% absolute reduction was seen mostly in cancellation of the order and predominately in housestaff clinicians (70% of the distribution). Interestingly, patient female gender (38% vs. 58%; p=0.02) and improving renal function were associated with lower alert compliance rates. In summary, the 226 alerts received by housestaff resulted in only a 42% compliance rate.
This study showed how a “real-time” pop-up CDS alert could reduce the number of medication errors and highlighted the ongoing issue of noncompliance with alerts. Furthermore, while more experienced housestaff were more likely to follow the alerts, it would be interesting to further elucidate the rationale for ignoring and/or overriding the alerts.
Second Article Review
This is a second review of the article titled, "A Trial of Automated Decision Support Alerts for Contraindicated Medications Using Computerized Physician Order Entry".
As past studies have shown, automated clinical decision support systems (CDSS) integrated with computerized order entry CPOE in clinical information systems can reduce medication errors and prevent adverse drug events, however clinicians continue to resist acknowledging and applying these types alerts to their everyday workflow. As a trial basis, the authors studied a group of alerts associated with renal insufficiency to evaluate the utility of the alerts and analyze the factors that may cause housestaff to be noncompliant with alert recommendations.
The study was conducted at The University of Illinois Hospital and Medical Center. They employ Cerner Millennium  as their EMR system for CPOE and its CDS system (Discern Expert). To start, a group of newly alerts were created based on a new serum creatinine level, along with a normalized level based only on gender and age. The alerts were designed to trigger when a clinician placed an order for on the contraindicated drugs for a patient who's recent creatinine clearance was less than the corresponding safe level of the drug. The alert would notify the clinician of the most recent creatinine level and with a recommended "safe" for each drug. A four month control period was also included in the study to assist on how effective the alerts would be once they were implemented. All alerts were logged and recorded within the EMR during the study. 
Over the next 14 months, the study showed 323 alerts were triggered. During the 4 month control period, 87 situations were identified in which the alert would have been triggered if the alerts were active. For both periods, the drug most commonly chosen was metformin and the results did not differ from patient gender, age,or degree of renal dysfunction. Overall the likelihood that a patient would receive at least one dose of contraindicated medication decreased from 89% to 47% after the alerts were implemented. As for factors leading to noncompliance by housestaff, it was determined that patient gender was associated with their compliance rate. Compliance in female patients was lower than that in male patients and overall decreased with improving renal function.
The study proved to be successful. For after implementing the CDS alerts, it was clear that it decreased the likelihood of clinicians completing contraindicated orders and the administration of the medications. However noncompliance is still an issue that will always play a factor on the way healthcare organizations utilize and implement CDS alerts. which can assist clinicians in delivering the best patient care they can to their patients. As new CDS systems create new guidelines and practices, the healthcare industry should focus on creating new and innovating ways to educate clinicians on the importance of utilizing CSDs alerts to ensure the safety of the patients.
- Galanter WL, Didomenico RJ, Polikaitis A. A trial of automated decision support alerts for contraindicated medications using computerized physician order entry. J Am Med Inform Assoc. 2005 May-Jun;12(3):269-74. Epub 2005 Jan 31. http://www.ncbi.nlm.nih.gov/pubmed/15684124