Imaging Interpretation Discrepancies: "Danger, Will Robinson!"
Imaging Interpretation Discrepancies: "Danger, Will Robinson!"
By Bruce D. Janiak, MD, FACEP, FAAP, Professor of Emergency Medicine, Medical College of Georgia; and Richard J. Pawl, MD, JD, FACEP, Assistant Professor of Emergency Medicine, Medical College of Georgia.
An emergency physician (EP) diagnoses pneumonia in a 40-year-old male and treats appropriately. Later, the radiologist agrees with the pneumonia diagnosis, but suggests that "a small nodule in the left upper lobe should be evaluated with a chest CT." Somehow, that report never gets read by the ED staff, and the patient's cancer progresses until it is too late for effective treatment. A lawsuit is filed.
The systems used to assure quality control over emergency department (ED)-generated imaging studies will necessarily vary from facility to facility due to the variances inherent to different hospital systems. That reality has contributed to the absence of any one system being proposed as being more advantageous to another in the literature. To be sure, it is in the interest of every emergency medicine group and every affiliated radiology department to have a system that assures any errors (or reading variations) that may occur in the ED radiographs are captured and acted upon appropriately. The obvious goals are to provide optimal patient care and reduce any malpractice liability that could arise from such errors.
Several studies regarding the discrepancies between EPs' and radiologists' readings of films generated out of the ED are worth mentioning at the outset. Simon and Khan et al published a prospective study in 1996 that evaluated the concordance rate between the interpretations of plain films by pediatric EPs and the pediatric radiologists at an urban tertiary care children's emergency room at the same institution.1 They found that the radiographic interpretations by the EPs were generally in agreement with the radiologists, and when there were variances, no adverse outcomes occurred.2 A study published in 1998 by Lufkin and Smith et al showed that out of 9,599 sets of radiographs interpreted by EPs in the study, there were only 11 discordant interpretations with the radiologists.3 Overall, EP interpretations of plain radiographs are usually concordant with the subsequent radiologist interpretations for the clinically significant aspects of the plain films. However, EPs are generally not held responsible for identifying all the subtleties of a plain radiograph as is a radiologist. For example, this may include the identification of pathological findings that may be pertinent to the patient's well-being, such as the incidental finding of a potentially cancerous lesion on a chest film, but do not affect the immediate care of that patient while in the ED. Furthermore, EPs and radiologists are still subject to human error, making them vulnerable to overlooking radiographic findings more urgently pertinent to the care of the ED patient, such as a small pneumothorax. Hence, a quality control mechanism that attempts to capture such errors and correct or attenuate any clinical consequences to the patient is a necessity.
Even though EPs would not be held to the standard of care of a radiologist if a cancerous lesion is missed on a plain film, the EP generally owns the responsibility of notifying their mutual patient of such a finding. Even though it is arguable that the radiologist and EP share that responsibility to notify the patient, the burden generally weighs more heavily on the ED group as the providers who have had closer contact with the patient. While there are scores of variations of the consequences of erroneously interpreted plain films, the point is that a system is needed to capture such errors on a timely basis and to act appropriately to prevent patient harm and malpractice liability.
X-ray discrepancy review systems, even when they function properly, can still lead to patient harm and malpractice liability. An illustrative case is Dustin Eastbourne v. Encino Tarzana Regional Medical Center, wherein the plaintiff was injured in a motorcycle accident in August 2004 (B203954, Cal. App. 11/20/2008). The plaintiff's injuries included a hand injury for which hand x-rays were performed. In the hospital's ED, the plaintiff was treated by an EP who failed to recognize a fifth metacarpal fracture on the x-ray. The following day, the x-rays were reviewed by a radiologist who reported the metacarpal fracture to the ED as per the departmental x-ray discrepancy system. A different EP called Mr. Eastbourne's home and left a recorded message. That message was reportedly never received, and the plaintiff, a carpenter, subsequently suffered continued pain from the injury and sought a second opinion. The subsequent orthopedic physician identified a fifth metacarpal malunion from the missed fracture. After conservative treatment failed, the respondent had to undergo operative reduction and fixation. Despite a good surgical outcome, the plaintiff reported recurrent pain and impairment of function that affected his ability to perform his duties as a carpenter. One issue at trial was whether the standard of care was met when the second EP called the patient and merely left a message on the patient's answering machine, an act that did not verify that the patient had received the message. One of the experts at trial opined that although there existed no written guidelines regarding the standard of care on this point, the expert would have made an attempt to assure that the patient had, in fact, been contacted rather than merely leaving a message. Without going into the legal details at issue when the case was reviewed by the appellate court, that court decided that there was sufficient evidence at the trial level to assert that the standard of care was not met by merely leaving a message on the plaintiff's message machine. Losing this issue contributed to the appeals court's decision to let the trial court's result stand, which had awarded over $43,000 to the plaintiff.
In 2000, Espinosa and Nolan published a longitudinal study of a system designed to reduce the errors made by EPs and the consequences of such errors at the hospital of one of the authors.4 The study showed the impact that cooperation between the radiology department and the ED can have a positive effect on reducing errors in reading radiographs and reducing the consequences to the patients whose films were erroneously interpreted.5 The system developed used Nolan's suggestions for designing a safe system of care.6 Nolan proposed that designers of systems of care can create safer systems by attending to three tasks:
"[D]esigning the system to prevent errors; designing procedures to make errors visible when they do occur so that they may be intercepted; and designing procedures for mitigating the adverse effects of errors when they are not detected and intercepted."7
Nolan's essay goes on to describe the ramifications of these tasks in applying them to systems of care. Espinosa's study exemplifies Nolan's guidelines, as they were applied to a quality assurance system meant to reduce the errors and consequences resulting from discordant interpretations of plain radiographs arising from an ED. Espinosa's system changes for ED plain x-rays included a design that was to be followed regardless of the time of day or day of the week. Previously, the facility had used four separate processes for handling plain-film interpretation depending upon the time of the day and the day of the week at which the film had been ordered. Additionally, the system clearly assigned primary responsibility for initial x-ray interpretations to the EP. A form for handling discrepancies was developed to provide feedback to the physicians and the process for handling the form was clearly defined and implemented. Finally, open communication and cooperation between the physicians of both specialties was encouraged.
After having innumerable conversations regarding this topic, it appears to the authors that the majority of emergency plain x-ray interpretations are initially provided by EPs. Some hospital radiology services provide "contemporaneous" readings by the radiologist during normal business hours (or through the evening), with readings defaulting to the EP after hours. Reports on more sophisticated imaging studies (e.g., computed tomography [CT] scans, magnetic resonance images [MRIs], ultrasound) are usually provided by the radiology service that is on call. Although the "significant error" rate for reading plain radiographs by trained and experienced EPs seems to be equal to that of radiologists, the systems for identification of these discrepancies and the action plan for dealing with them have not been standardized. The errors resulting from the system failures can be categorized:
Failure to inform the radiologist of the EP's interpretation. This is an issue of compliance. Even the most obsessive EP may omit his or her interpretation, forcing the radiologist to assume that the finding was missed, resulting in an unnecessary phone call or message to the ED.
Failure of the radiologist to inform the EP of his/her reading. This is an uncommon error and it usually takes the form of an incomplete initial report ("wet read") followed by a later, more complete rendition which results in a clinical change.
Example: Initial report –"No acute process" final report –"Suspicious right upper lobe lesion, suggest CT." The final report is received in the ED the next morning, long after the patient has been discharged.
A common variation on the above is an initial reading of "pneumonia" followed by the "suggest CT" report. The resolution of these issues requires a tight discrepancy review system in the ED and accurate chart demographics to aide in communicating with the patient.
Failure of the EP to correctly interpret the initial radiology report in clinical context. Example: 13-year-old male with testicular pain; ultrasound report: "no torsion, decreased flow." In this case, the EP focused on "no torsion," forgetting that "decreased flow" was a red flag for vascular compromise. The patient was discharged and a lawsuit was subsequently filed for failure to diagnose testicular torsion. This type of error is virtually impossible to prevent by any type of radiological call-back system due to the need for urgent action to prevent the adverse outcome.
Failure of the ED system to identify significant discrepancies and act upon them by notifying the patient. Example: Official radiology report suggests a CT scan for a nodule noted in the lung. The report arrives in the ED the next morning, but the position assigned to screen the reports (e.g,. the nurse practitioner) is ill. The ED is very busy, and the reports are set aside by the clerk. The next day, many of the reports have been discarded by someone who thought they had been reviewed. The lung cancer is discovered six months later, too late to save the patient.
Failure of the ED to have a definitive policy. Example: As above, a "position" is assigned to review the x-ray reports. It could be a mid-level provider, the charge nurse, a resident, or an attending. Any provider could become too busy with current patients to spend time with the review process. This quality review is too important to delegate to someone "if they have the time." A tighter and better-defined system is needed.
Failure of radiology to notify the primary care provider (PCP). Example: The ED system above "misses" the "suggest CT for lung nodule" report. As instructed, the patient, thinking they have a simple pneumonia, follows up with their PCP one week after the ED visit. Either the PCP has missed the report, filed the report without reading, or the radiology department never faxed or mailed the report. The lung cancer continues undiagnosed.
Failure of the EP to review the official report after they have provided a partially correct interpretation. Example: 45-year-old male with right lower quadrant pain clinically consistent with kidney stone. The EP orders CT for a "stone hunt" and is correct that no stone can be seen. Official report agrees and adds the diagnosis of appendicitis. The patient returns with a ruptured appendix. A lawsuit follows.
Failure of the EP to review the official report after he or she has incorrectly interpreted the image. Example: As outlined above, but a ureteral stone was diagnosed by the EP. Unfortunately, the stone is actually a phlebolith and the diverticulitis noted by the radiologist remained undiagnosed and untreated. The reading discrepancy was never noted or acted on by the ED's system for reviewing reports. The patient experienced a ruptured diverticulum and died.
The examples above represent primarily errors in communication between providers. Although no system can be perfect, we offer some suggestions on possible ways to make the process of imaging discrepancy review safer:
Re-program the picture archiving and communication system (PACS) to require both the radiologist and the EP to add their interpretation to the screen. Failure to "fill in the blank" will freeze the screen until the physician completes the requirement.
Tighten up the ED review policy. Assign reviews to a specific individual each day and require check-off on completion.
Formalize the process when discrepancies are identified. For example, the screener gets the input of the emergency resident or attending and develops a written plan of action. (This can be accomplished with simple forms). Most discrepancies are minor and fit into the "no action required" category. Forms should have a place to record results of a phone call to the patient or patient's PCP. Patients with disconnected phones may require a certified letter. All completed action plan forms should become a part of the patient's chart.
Assure that the system used is applicable to all hours of the day and all days of the week, avoiding the confusion that occurs when multiple systems are used.
Summary
There is no system of care protocol that can prevent all potential errors, but creating a radiological call-back system to control for errors that occur is an essential aspect of providing emergency care. Although systems may vary from site to site, effective call-back systems should be simple and broadly applicable systems across all hours of the day and all days of the week. The providers involved in implementing the system should have clearly defined responsibilities and firm accountability for the system. The ED and the radiology providers should work cooperatively so that the system works smoothly for both departments. And finally, the system should promote ongoing, open communication amongst the providers of both departments to allow for flexibility in problem-solving as unanticipated adverse events occur.
References
1. Simon HK, Khan NS, Nordenberg DF, et al. Pediatric emergency physicians' interpretation of plain radiographs: Is routine review by a radiologist necessary and cost-effective? Ann Emerg Med 1996;27:202.
2. Id. The study was small in number of radiographs reviewed and the reviewed time interval was only one month, so there are substantial limitations of the study, but the study is consistent in their findings with other studies regarding emergency physician/radiologist reading concordance with plain films.
3. Lufkin KC, Smith SW, Matticks CA, et al. Radiologists' review of Radiographs interpreted confidently by emergency physicians infrequently leads to changes in patient management, Ann Emerg Med 1998;31:295.
4. Espinosa JA, Nolan TW, Reducing errors made by emergency physicians in interpreting radiographs: A longitudinal study, BMJ 2000;320:737-740.
5. Id.
6. Nolan TW, System changes to improve patient safety, BMJ 2000;320:771-773.
7. Id. at 771.
An emergency physician (EP) diagnoses pneumonia in a 40-year-old male and treats appropriately. Later, the radiologist agrees with the pneumonia diagnosis, but suggests that "a small nodule in the left upper lobe should be evaluated with a chest CT." Somehow, that report never gets read by the ED staff, and the patient's cancer progresses until it is too late for effective treatment. A lawsuit is filed.Subscribe Now for Access
You have reached your article limit for the month. We hope you found our articles both enjoyable and insightful. For information on new subscriptions, product trials, alternative billing arrangements or group and site discounts please call 800-688-2421. We look forward to having you as a long-term member of the Relias Media community.