Electronic requests for imaging studies from clinicians frequently include erroneous clinical information, a problem that may lead to a variety of troubling issues, challenges, and choppy waters for radiology departments, researchers from the Republic of Ireland have found.
In a retrospective study conducted over a six-month period during 2013, a group from Cork University Hospital found that nearly half of imaging requests included incorrect biochemical and/or hematological information.
"This may result in inappropriate prioritization, protocoling, and administration of IV contrast," said staff radiologist Dr. Maria Twomey, who shared the research during a scientific session at the RSNA 2014 meeting in Chicago. "This also impacts the accuracy of the consequent radiology report."
A level I trauma center and a tertiary cancer referral center, Cork University Hospital performs approximately 270,000 radiology exams each year. The institution uses a RIS and PACS (Agfa HealthCare); radiologists can see electronic imaging requests on the PACS during reporting, but there are no electronic patient records, and the RIS imaging request system is not directly linked to the laboratory system, she noted.
Anecdotal evidence
Experiencing anecdotal evidence of incorrect clinical information on imaging requests, the researchers suspected that clinical colleagues were entering erroneous data to get scans performed faster.
"I wanted to formally assess and quantify this, because accurate clinical information is paramount for the radiologist to accurately prioritize, protocol, and report an imaging study," Twomey said.
In their study, the researchers sought to determine the rate of erroneous biochemical and hematological parameters included on electronic requests for CT pulmonary angiography (CTPA), CT thorax, and abdominopelvic CT exams. They retrospectively selected a random sample of 250 electronic imaging requests submitted on their institution's RIS between July and December 2013. The study sample included 100 CTPAs; 70 CT exams of the thorax, abdomen, and pelvis; and 80 abdominopelvic CT exams.
They then compared the laboratory levels supplied by the referring clinician to the actual reported levels on the biochemical and hematology electronic reporting system. In a subgroup of patients who received CTPA studies, the researchers also compared D-dimer and pO2 levels.
Of the 250 image requests, 45% had incorrect biochemical and/or hematological information, according to the researchers. Delving further into the data, they found a significant number of incorrect D-dimer and pO2 levels in the CTPA requests, and 15% reported an abnormal D-dimer result, although the actual reported levels were normal. In addition, 25% of the requests noted hypoxia, even though the actual reported pO2 level was normal.
"Ninety percent of the incorrect lab results were entered by internal medicine physicians," Twomey said. "This was statistically significant and was mostly entered by internees and junior residents."
An inaccurate creatinine level was included in 10% of all requests. Furthermore, 70% had incorrect iron deficiency anemia requests. Elevated C-reactive protein and/or white cell counts were reported in 70% of acute abdominopelvic CT requests. However, 20% of the actual results in this group were normal.
The researchers also noted that only 15 of the 250 electronic imaging requests were completed by consultants (senior doctors); four of these had inaccurate information. Interns (trainees) contributed the largest percentage (45%) of imaging requests.
Medical physician referrals had a significantly higher incidence of erroneous lab results when compared with surgeons, Twomey added. The difference was statistically significant (p = 0.01).
A cause for concern
The high level of erroneous clinical and laboratory information is concerning, Twomey continued.
"The primary concern is the effect that this can and does have on how the radiologist protocols, prioritizes, and reports the study," she said.
For example, the incorrect modality or study protocol could be performed and patients could receive an inappropriate radiation dose. It could also lead to a delay in other patients being scanned and may contribute to misinterpretation of radiological findings in the setting of incorrect or misleading clinical information, according to Twomey.
In an ideal world, all biochemical and hematological results should be double-checked with the laboratory system. However, this would be laborious and affect workflow.
The study findings support the implementation and use of an electronic ordering system that would be linked directly to a patient's laboratory and, ideally, electronic patient chart. While this software exists and is in use, it's not currently available at their and many other institutions, Twomey said.
Budgetary constraints are omnipotent throughout radiology, but these findings would support investing in this type of software, she said, also emphasizing that clinicians must be made aware that providing incorrect information is potentially deleterious to patient management and does not foster productive professional colleague interaction.
