Radiology Errors Are the Worst, USP Says

Cheryl A. Thompson

Somewhere in the United States a patient died after health care personnel dispensed an ionic contrast medium and administered an intrathecal injection of it, rather than the called-for nonionic contrast agent, during a radiology exam. Elsewhere, a patient with abnormal renal function died from complications of gastrointestinal-tract bleeding after someone in the cardiac catheterization laboratory obtained eptifibatide from the unit's automated dispensing device and administered a full, unadjusted dose of the platelet aggregation inhibitor.

These two events represent perhaps the worst of the 2032 medication errors associated with radiological services at hospitals and health systems that anonymously submitted reports in 2000–04 to a United States Pharmacopeia (USP) database.

Twelve percent of the errors resulted in harm to patients—seven times the overall frequency at which errors during the five-year period harmed patients, USP announced January 18 in its sixth annual data report from the Medmarx program.

"The 12% is also the highest percentage of harm compared to any previously published report of Medmarx data," primary author John P. Santell, with the USP Center for the Advancement of Patient Safety (CAPS), told reporters during a media briefing. "We believe this finding alone is a signal that medication error involving radiological services is a serious issue, especially given the large number of radiologic exams, approximately 60 million [annually], which involve the use of a radiologic drug."

Dissent from the radiology profession. About the time Santell spoke, the American College of Radiology distributed a press release calling USP's report an "incomplete, inaccurate" description of errors in medical imaging facilities.

To buttress its case, the college reported what it called the medication-error rate for radiological services—calculated by dividing the 2032 errors reported anonymously to the Medmarx database by 315 hospitals and health systems by the 2.5 billion imaging procedures conducted nationwide in 2000–04. This so-called error rate of 0.00008%, the college said, is more than 3700 times smaller than the lowest hospitalwide medication-error rate mentioned in the Institute of Medicine's seminal report To Err Is Human.

The college also challenged USP's lack of detail when identifying the staff responsible for initially making the errors: USP lumped radiologists with other physicians.

Of the 2032 errors, 825 were associated with the cardiac catheterization laboratory, where, according to the college, less than 1% of the procedures are performed by radiologists.

The college represents radiologists, radiation oncologists, interventional radiologists, nuclear medicine physicians, and medical physicists but not the medical specialists who predominate in the nation's cardiac catheterization laboratories.

Further, the college bristled at radiological personnel taking the blame for events—such as injections ordered by physicians in other hospital departments—unrelated to imaging procedures but nonetheless occurring or continuing in a radiology service area.

Radiological services as an island. Clinical pharmacist Thomas J. Barrs, who has studied the imaging field since 1995, said the higher-than-average frequency of harm from radiology-connected medication errors is likely a consequence of the way radiological services evolved in hospitals.

"I think from the beginning, radiology was envisioned by people not in the field . . . as something very esoteric, almost magical," said Barrs, who works at 481-bed DeKalb Medical Center in Decatur, Georgia.

Over time, radiological services evolved independently of their host institutions, "almost like an island," he said.

But as the field of imaging diagnostics grew and the level of care needed by inpatients intensified, Barrs said, radiological service areas began encountering patients who were receiving complicated drug regimens in progress, such as heparin infusions controlled by pumps, that were unrelated to a diagnostic procedure.

"The personnel in imaging departments simply aren't accustomed to that," he said of the complicated regimens, "and unfortunately, for one reason or another, no one ever really turned their attention to that area" until recently.

Barrs disputed the radiology association's argument that radiology services staff should not bear responsibility for errors involving medications ordered by other departments.

"When a patient comes unattended to any hospital department for some kind of procedure that involves the administration of a drug through existing [intravenous] lines . . . they're yours," he said. "I mean it doesn't really matter what the patient's got going, they become your responsibility at that point."

Hospitals' most harmful errors. USP in July 2004 used its USP Patient Safety CAPSLink to report briefly on five years' worth of errors connected to radiology departments and cardiac catheterization laboratories but not nuclear medicine. At that time, 16.3% of the errors linked to radiological services had harmed patients—eight times the overall frequency of harm for errors in the database.

Errors in radiological service areas in the new report even surpassed the propensity for harm of medication errors in intensive care units, where presumably all the patients are critically ill and perhaps not able to rebound readily from mistakes.

Barrs, who works about half of the time in an intensive care unit, attributed the difference between the two service areas' level of error-related harm to the presence or absence of pharmacists.

Personnel in the radiology department usually do not have access to pharmacists, he said, but nurses in the intensive care unit constantly interact with clinical pharmacists.

But are those substances really medications? Physical absence aside, pharmacists prospectively review relatively few of the orders for medications to be administered in radiological service areas.

According to the 2005 ASHP national survey of pharmacy practice in hospital settings, 8.5% of the respondents' facilities had a formal policy requiring pharmacists' review and approval of medication orders before the drugs' administration in radiology, and 5.6% had a similar policy for the cardiac catheterization laboratory.

That survey was conducted the year after the Joint Commission on Accreditation of Healthcare Organizations (JCAHO) added diagnostic and contrast agents and radioactive medications to its definition of medications.

Because of that addition, pharmacists in JCAHO-accredited hospitals are supposed to review the medication orders for radiology services before the agents are obtained from floor stock or an automated storage-and-distribution device or dispensed by the pharmacy. JCAHO said it allows exceptions when a licensed independent practitioner, such as a radiologist, controls the agents' dispensing and administration.

Barrs and Santell said they disapprove of JCAHO's decision last June to also exempt oral contrast media not stored in the pharmacy.

Oral contrast media accounted for less than 1% of the harmful radiology-associated errors in the Medmarx database for 2000–04 and about 3.7% of the radiology errors that did not harm patients.

Action needed, USP says. Many of the radiology errors reported to the database, Santell said, resulted from a breakdown in continuity of care, communication, the transfer of important medical information, or a facility's process or procedure.

Hospitals, health systems, and health care providers need to examine their medication processes for radiological services, identify the weaknesses, and strengthen those areas in order to reduce the risk of errors, he said.

Copies of MEDMARX Data Report: A Chartbook of 2000–2004 Findings from Intensive Care Units and Radiological Services can be purchased from the USP Web site, www.usp.org.