5 ways feedback improves rad performance

The Institute of Medicine’s 2015 report Improving Diagnosis in Health Care revealed the persistence of diagnostic errors in all settings of healthcare that continued to harm an unacceptable number of patients. Numerous factors contribute to hindering the diagnostic process, such as workflow limitations, poor implementation of technology, the medical liability system, current reimbursement models and organizational culture. Nonetheless, the authors conclude that improving the diagnostic process is “not only possible, but also represents a moral, professional and public health imperative.” 

Radiology plays an important role in the diagnostic process. Radiologists utilize their knowledge of imaging modalities, imaging protocols, image interpretation and clinical management recommendations to advise other healthcare professionals on diagnosis and management. Radiologists can improve their performance by obtaining feedback from referring clinicians and colleagues. This feedback loop can provide actionable information to adjust future clinical performance to improve patient care.

According to the author of a chapter in the ebook Learning Theories in Plain English (Vol. 1 of 2), the concept of using feedback loops to maintain control of processes and systems has existed since the 18th century. Through observations of the natural world, we find that organisms commonly use feedback loops to maintain equilibrium or change existing states. Similarly, his concept can be applied to improving the performance of radiologists. By understanding the outcomes of our actions, we can either retain those practices that have desirable outcomes, or we can change them to improve outcomes. The key is to provide the feedback so that radiologists can learn and continue to improve.

Here are five specific opportunities to obtain feedback that can improve radiologist performance and increase the value of radiology services: 

1. Standardized Reporting

The radiology report represents the primary means of communication between the radiologist and the referring clinician. Using consistent structure and language can improve communication, comprehension and completeness. Increasingly, more specific information is being requested by referring clinicians to assist in patient-care decisions. Feedback from referring clinicians can be utilized to improve the value of radiology reports by including pertinent data that impacts clinical decision-making. For example, a clinician on the liver transplant team needs to know whether a patient has hepatocellular carcinoma and, if so, the size of these lesions. This information will determine treatment options and priority on the transplant list. However, the diagnosis of hepatocellular carcinoma can be difficult on imaging studies. Consequently, initiatives such as standardized reporting of liver findings on imaging examinations using the American College of Radiology’s Liver Imaging Reporting and Data System (LI-RADS) are being promoted to standardize terminology, reduce image interpretation variability and enhance communication with referring clinicians. Ideally, the radiologists and referring clinicians should come to an agreement on the reporting standards to maximize the utility of information available from imaging examinations.

2. Collaborative Management Recommendations

Incidental findings are commonly encountered in imaging examinations. While most of these findings are benign, unnecessary imaging or intervention is often performed to avoid missing the early presentation of potentially treatable disease. For some incidental findings, such as pulmonary nodules, standardized management recommendations exist, in the form of Fleischner Society recommendations, and are widely accepted. For other incidental findings, such as thyroid nodules, multiple management recommendations from different organizations cause confusion and disagreement. Collaboration with your referring clinicians can be helpful to standardize the care in your local community. Feedback from referring clinicians identifying which recommendations to follow can reduce the variability of radiologist recommendations and strengthen the confidence of radiologists in adhering to standardized management recommendations. Referring clinicians also can gain confidence in adhering to radiologist recommendations that have been developed in collaboration with the relevant subject matter experts in the local community. At NYU Langone Health, the radiology department collaborated with the subject matter experts at the institution to collectively endorse a set of management recommendations for several incidental findings. For example, radiologists collaborated with vascular surgeons to jointly endorse management recommendations for incidental aortic aneurysms. Similarly, standardized collaborative management recommendations for pulmonary nodules, ovarian cysts, pancreatic cysts and renal cysts were collectively endorsed.

3. Radiology-Pathology Correlation

Pathological correlation of imaging findings is a valuable source of feedback to improve radiologist performance. Such feedback enables the radiologist to assess her clinical reasoning process and hone her diagnostic strategy for future cases. For example, if a radiologist continues to diagnose appendicitis and pathology often discovers no evidence of inflammation, the feedback provides the motivation for the radiologist to alter her diagnostic strategy. As opposed to peer review, pathological correlation offers more clarity and less emotional impact for the interpreting radiologist. Nevertheless, obtaining pathological correlation can be a time-consuming and inefficient process. With the advent of integrated EMRs, the pertinent data exist in a single system ready to be searched and sorted. At NYU Langone Health, we have implemented algorithms to automatically match radiology reports with pertinent pathology reports and automatically deliver these correlations to the interpreting radiologist. This timely feedback serves as a learning opportunity for the individual radiologist. Aggregated feedback from multiple radiologists can be shared in peer learning conferences so that additional radiologists can benefit. Additionally, these cases can provide the impetus to perform clinical research and improve diagnostic interpretations on a larger scale.

4. Radiology-Operative Note Correlation

For certain radiology subspecialties, such as musculoskeletal and cardiovascular imaging, pathological correlation is often sparse. In these subspecialties, operative procedures, such as arthroscopy and angiography, are more commonly performed. The findings on these operative procedures can provide equally valuable feedback on radiologist interpretations. Similar to pathology reports, the operative reports that exist in the EMR can be matched to radiology reports and delivered to the radiologist. In this manner, clinical feedback can be provided for a larger number of cases and for more radiologists. 

5. Follow-up Imaging Correlation

Beyond pathology and operative reports, follow-up imaging provides another opportunity to obtain feedback on interpretations. The radiology reports for these follow-up examinations exist in the EMR and can be readily matched to the earlier reports. For example, a liver lesion is seen on an ultrasound examination and an MRI is recommended for further evaluation. If the MRI is subsequently performed, the MRI report can be delivered to the radiologist who interpreted the ultrasound examination. This feedback can help the radiologist to evaluate the outcome and appropriateness of her recommendation. Another example might be a lung lesion diagnosed as rounded atelectasis by the radiologist. If a follow-up examination is performed, it would be valuable for the initial radiologist to have confirmation of her diagnosis or evidence of an alternative diagnosis.

Feedback can be an effective tool to improve radiologist performance. Numerous opportunities exist to collect feedback from various sources. Some require collaboration with referring clinicians while others require investments in information technology. While it may require some effort to obtain this feedback, it will be worthwhile if it can help radiologists to generate more valuable reports, provide standardized management recommendations and improve their interpretations. 

Dr. Kim is an associate professor and the director of quality and safety in the radiology department at NYU Langone Health.