Radiologic diagnostic errors can result in delayed or unnecessary treatment. This paper describes five types of cognitive biases, with accompanying case study and images, that result in misdiagnosis. Both individual education and system-level solutions are described.

Changes in healthcare delivery due to the COVID-19 pandemic resulted in delays in care that can lead to patient harm. In this study using patient safety event data submitted to the VHA National Center of Patient Safety, researchers identified healthcare delays involving laboratory results, treatment and interventional procedures, and diagnosis.   

Previous research has identified racial and ethnic discrepancies in pulse oximetry measurement which can lead to delays in diagnosis or treatment. This editorial discusses racial and ethnic biases in clinical algorithms and devices and two emerging approaches – photoacoustic imaging and polarized light oximetry – that have potential to address the racial and ethnic biases in pulse oximetry.

Misinterpretation of radiology test results can contribute to diagnostic errors and patient harm. Using a set of 838 pediatric and adult radiographic examinations, this retrospective study found that radiographers’ interpretations can complement emergency clinicians’ interpretations and increase accuracy compared to emergency clinician interpretation alone.

The effective use of resources through stewardship initiatives can support error reduction through focusing actions of care. This commentary discusses how diagnostic stewardship can enhance diagnostic testing behaviors across the diagnostic process.

Missed diagnosis of infectious diseases can have serious consequences for patient safety. This article describes a delayed diagnosis of disseminated tuberculosis in a patient with lung cancer and discusses the how cognitive biases and systems failures contributed to the diagnostic error.

Metacognition (e.g., when one reflects on one’s own decision and decision making) is an approach to reducing diagnostic errors. Using data from the Melanoma Pathology Study (M-PATH) and Breast Pathology Study (B-PATH), researchers assed pathologists’ metacognition by examining their diagnostic accuracy and self-confidence. Results showed pathologists with increased metacognition sensitivity were more likely to request a second opinion for incorrect diagnosis than they were for a correct diagnosis.

Studies have shown diagnostic discordance in evaluation of surgical pathology specimens. In this study, pathologists and pathology residents were asked to diagnose surgical pathology or cytopathology cases and provide a diagnosis. Most respondents provided the correct diagnosis for most of the cases; 35% of cases were wholly or partially misdiagnosed. Educational and process changes (e.g., requiring subspecialist over-read for some diagnoses) were implemented in the pathology department in response, resulting in substantial improvement in error rates.

Wrong blood in tube (WBIT) errors can result in serious diagnostic and treatment errors, but may go unrecognized by clinical staff. In this study, machine learning was used to identify potential WBIT errors which were then compared to manual review by laboratory staff. The machine learning models showed higher accuracy, sensitivity, and specificity compared to manual review. 

Black and brown patients have experienced disproportionately poorer outcomes from COVID-19 infection as compared with white patients. This study found that patients who identified as Asian, Black, or Hispanic may not have received timely diagnosis or treatment due to inaccurately measured pulse oximetry (SpO2). These inaccuracies and discrepancies should be considered in COVID outcome research as well as other respiratory illnesses that rely on SpO2 measurement for treatment.

Diagnostic errors in radiology can result in treatment delays and contribute to patient harm. This article provides an overview of the common cognitive biases encountered in diagnostic radiology that can contribute to diagnostic error, and strategies to avoid these biases, such as the use of a cognitive bias mitigation strategy checklist, peer feedback, promoting a just culture, and technology approaches including artificial intelligence (AI).

Root cause analysis is a commonly used tool to identify systems-related factors that contributed to an adverse event. This study assessed a system-based approach, (i.e., collaborative case reviews (CCR) co-led by radiology and an institutional patient safety program) to identify contributing factors and explore the strength of recommended actions in the radiology department at a large academic medical center. Stronger action items, such as standardization of processes, were implemented in 41% of events, and radiology had higher completion rates than other hospital departments.

Teleradiology - general radiologists who support several hospitals and read films remotely – can increase off-hours coverage but this approach can result in increased errors. This retrospective review examined errors and discrepancies between teleradiology findings and image interpretation from local facility radiologists. Most errors involved CT scans; the most common errors included missed fractures or dislocations and bleeding.

A shared understanding of terminology is critical to providing appropriate treatment and care. This study assessed pathologist and clinician agreement of commonly-used phrases used to describe diagnostic uncertainty in surgical pathology reports. Phrases with the strongest agreement in meaning were “diagnostic of” and “consistent with”. “Suspicious for” and “compatible with” had the weakest agreement. Standardized diagnostic terms may improve communication.

Artificial intelligence (AI) has been used and studied in multiple healthcare processes, including detecting patient deterioration and surgical decision making. This literature review focuses on studies using AI to detect breast cancer in mammography screening practice. The authors recommend additional prospective studies before using artificial intelligence in clinical practice. 

Misdiagnosis of cardiovascular conditions can lead to serious patient harm. This systematic review found that misdiagnosis of acute myocardial infarction (AMI) occurs in approximately 1-2% of cases, and AMI is commonly diagnosed as other heart conditions, musculoskeletal pain, or gastrointestinal disease. The authors suggest that there are opportunities to reduce cases of missed AMI with better education about atypical symptoms and improved training of electrocardiogram interpretation.