CODEX Digest - 10.2.25

Diagnostic excellence: turning to diagnostic performance improvement. (subscription required) 

Auerbach A, Raffel K, Rasooly IR, et al. Diagnosis (Berl). Epub 2025 Sep 16 

Organizational commitment to diagnostic improvement is necessary to measure error and achieve excellence. This commentary reviews the importance of Diagnostic Excellence Programs, shifting from an epidemiologic approach to targeting overall performance. System-focused initiatives, individual clinician behavior enhancements, patient engagement strategies, and cautious AI implementation are recommended for diagnostic excellence program development. 

Recognizing atrial flutter in the emergency department: challenges in diagnosis. (subscription required) 

Avidan Y, Aker A, Sliman H, et al. Am J Emerg Med. 2025;96:224-229.

ECG misinterpretation can lead to cardiovascular disorder misdiagnoses. This retrospective Israeli single-center analysis looked at 2,003 emergency department ECGs of patients discharged with atrial fibrillation (AF) or atrial flutter (AFL) diagnoses to assess accuracy of non-computer-aided conclusions. Results showed 44% of the 209 AFL cases to be misdiagnosed as AF cases. The authors found factors associated with misdiagnosis to be older age and prior AF diagnosis and call for targeted education rather than improving computerized accuracy.  

Exploring the risks of over-reliance on AI in diagnostic pathology. What lessons can be learned to support the training of young pathologists? 

Bellahsen-Harrar Y, Lubrano M, Lépine C, et al. PLoS ONE. 2025;20(8):e0323270.  

Independent pathologist judgement is a valuable skill when using AI for slide assessments. This French study contrasted the diagnostic accuracy of eight pathologists with varying levels of experience in cases with and without AI assistance. AI-supported pathologists achieved higher accuracy, but there was some evidence of overreliance and automation bias with AI in less-experienced pathologists who ignored lower-confidence ratings coming from the AI model.

Implementation of a quality improvement and clinical decision support tool for cancer diagnosis in primary care: process evaluation.

Chima S, Hunter B, Martinez-Gutierrez J, et al. JMIR Cancer. 2025;11:e65461. 

Clinical decision support (CDS) is an important contributor to diagnostic excellence. This mixed-method analysis surfaced successes and barriers to the use of Future Health Today, a CDS tool with quality improvement support, supporting timely diagnosis of cancer in 21 Australian primary care practices. Uptake of the tool was low with variable numbers of patients flagged for cancer. The results indicate that time, resources, individual practice characteristics, and cumbersome application support were barriers to implementation of a primary care CDS tool. 

In defence of causing patients to worry: ethical issues in the communication of diagnostic uncertainty.

Cox C, Fritz Z. Bioethics. 2025;39(7):700-708.  

Finding the balance between transparency and causing distress can be challenging when communicating diagnostic uncertainty. This commentary argues for the importance of disclosing uncertainty despite worry it might cause patients. The authors discuss evidence supporting such transparency and provide recommendations to assist physicians in communicating uncertainty to minimize worry. 

From cases to confidence: developing diagnostic reasoning skills through collaborative learning in graduate nursing education.

Jackson ML. Nurs Educ Perspect. 2025;46(5):319-321.  

The development of diagnostic reasoning skills in nurses supports diagnostic teamwork. This commentary highlights an online teaching strategy that employs AI-generated scenarios in collaborative learning to train first-year advanced practice students in diagnostic reasoning. Seventeen students participated, and post-program ratings showed increased confidence in diagnostic assessment domains.

Multicancer detection tests for screening: a systematic review. (subscription required)  

Kahwati LC, Avenarius M, Brouwer L, et al. Ann Intern Med. Epub 2025 Sep 16.

A proactive, collective method for early disease identification in asymptomatic adults could change cancer care. This systematic review examined the current state of multicancer detection (MCD) tests to efficiently diagnose a variety of cancers. No studies looked at benefits of screening, accuracy was mixed, and one study found harms. MCD tests are not ready for prime time.  

The self-report fallacy: when diagnosis predominantly relies on subjective symptom report. (subscription required) 

Merten T. Curr Opin Psychol. 2025;65:102096. 

Patient sharing of symptoms is a primary, yet imperfect, initiator of the diagnostic process. This opinion piece provides a short review on the value and potential harms associated with patient self-report in mental healthcare. The author highlights cognitive biases that affect physician interpretation of reported symptoms and makes the recommendation that the diagnostic process for mental-health diagnoses use objective data to ensure effectiveness and accuracy. 

Feasibility of pediatric diagnostic quality measurement in all united states hospitals. (subscription required) 

Michelson KA, Grubenhoff JA. Ann Emerg Med. Epub 2025 Sep 3.

Diagnostic measurement is necessary to understand the scope and improvement of misdiagnosis. This cross-sectional study of pediatric visits to US emergency departments looked at the volume of pediatric visits at high risk for misdiagnosis. They found volumes of most conditions were too low to reliably measure misdiagnosis in the majority of emergency rooms. Combining or unifying diagnostic measurement may be key to appropriate misdiagnosis tracking. 

Imaging-based surgical site infection detection using artificial intelligence. 

Muaddi H, Choudhary A, Lee F, et al. Ann Surg. 2025;282(3):419-428.  

AI shows promise in improving workload and accuracy for image triage. This cohort study used National Surgical Quality Improvement Program (NSQIP) data to identify patients that submitted surgical incision images to the patient portal for infection and to evaluate different AI tools for assessing the images. Although 93% of participants were White, one tool outperformed the other models for incision and infection detections. The assessment illustrates the potential value in using AI to review patient-submitted incision images to streamline image triage and facilitate earlier surgical site infections detection. 

Clinical, scientific and healthcare system consequences of misdiagnosing neonatal sepsis. (subscription required) 

Popescu CR, Lavoie PM. Expert Rev Anti Infect Ther. 2025;23(9):683-690. 

Sepsis is a consequential diagnosis and is especially dangerous for babies. This review looks at the consequences of misdiagnosing newborn sepsis. It explains the difficulties in correctly identifying cases and highlights studies showing that rates of sepsis vary widely across regions and health systems, particularly those in low-to-middle income countries. Detection differences occur due to a combination of inconsistent diagnostic definitions, limited access to reliable laboratory testing, overlap with non-septic conditions, and health system disparities.  

**Pediatric sepsis was featured in a recent webinar.   

Enhancing equity and efficiency in cervical screening uptake: a multidisciplinary quality improvement initiative. 

Santos C, Roye J, Tucker J, et al. BMJ Open Qual. 2025;14(3):e003111.     

Lack of uptake of screening programs contributes to unnecessary diagnostic delay and preventable harm. This quality improvement report discusses a British hospital’s initiative to reduce age-related cervical screening disparities in one community through patient-centered strategies that increase screening participation. The results increased screening for both age groups between 25-49 and 50-64, through a focus on building trust, improving accessibility, and patient-focused messaging.