High Rate of Overridden Drug Contraindication Alerts for Patients with Cancer

April 2014, Vol 5, No 3

Washington, DC—Patients with cancer are often prescribed medications for the treatment of comorbid conditions, and some of those medications may have harmful drug interactions with the therapies used to treat their cancer. Computerized alert systems are now being used to inform providers or support personnel on such contraindications.

A new study reported at the 2013 American Medical Informatics Association meeting suggests that the vast majority of contraindication alerts for prescribed cancer medications on computerized provider order entry systems are overridden by various providers, including physicians and nurses, and many of these are done without an explanation.

Of the 214 contraindicated medication alerts for a population of 117 patients with cancer (adults and children) who were being treated at a large, multispecialty, ambulatory academic practice that provides care throughout the Houston community, 93% of alerts were overridden, according to data gathered by Elise G. Brune and Dean F. Sittig, PhD, Professor, University of Texas Health Science Center at Houston, under the supervision of Allison B. McCoy, PhD, Assistant Professor of Biostatistics and Bioinformatics, Tulane University, New Orleans, LA. Furthermore, 83.2% of the alerts were overridden without an explanation.

Study Details
Physicians and nurses overrode all medication alerts they received, whereas medical assistants overrode 13% (8 of 62) of alerts. Members of several medical specialties overrode all alerts, including oncologists, obstetrician/gynecologists, endocrinologists, and neurologists.

By contrast, otolaryngologists did not override most alerts, nor did pediatricians and plastic surgeons. The sample sizes were too small to determine whether any of the differences in alert override rates were significant.

“The differences between specialty and provider types [of alerts] are likely due to those types with the highest override rates using the system more and seeing more alerts than others,” said Dr McCoy. “These high rates are not unique to cancer patients or our institution.”

On average, there were 1.7 overridden alerts per patient. Of the 46 alerts for absolute drug contraindications, 95% were overridden, as were 15.8% (12 of 76) of alerts for potential contraindications, and all 75 of the alerts for contraindication cautions.

The Rationale
Of the overrides with explanations, 5% were attributed to the benefit of the contraindicated medication outweighing the risk to the patient. Another 3% of the alerts were overridden because the clinician deemed the contraindication to present little risk to the patient, and the clinician gave appropriate warnings to the patient. A total of 2% of alerts were overridden because there were no reasonable alternatives for the patient. A total of 83.2% of the alerts were overridden without an explanation. Only 7% of alerts were not overridden.

The researchers also found peaks and valleys in the alert and override patterns: the highest rates of alert overrides were in March, May, September, and December, with the lowest rates in January, April, June, and July. The rates of overriding also peaked at 9:00 am and at 1:30 pm, and dipped overnight and at noon.

Dr McCoy has received a grant to evaluate all medication alerts in detail, not just those for cancer medications.

Related Articles