From second to hundredth opinion in medicine: A global consultation platform for physicians

Deep learning-aided decision support for diagnosis of skin disease across skin tones

Although advances in deep learning systems for image-based medical diagnosis demonstrate their potential to augment clinical decision-making, the effectiveness of physician-machine partnerships remains an open question, in part because physicians and algorithms are both susceptible to systematic errors, especially for diagnosis of underrepresented populations. Here we present results from a large-scale digital experiment involving board-certified dermatologists (n = 389) and primary-care physicians (n = 459) from 39 countries to evaluate the accuracy of diagnoses submitted by physicians in a store-and-forward teledermatology simulation. In this experiment, physicians were presented with 364 images spanning 46 skin diseases and asked to submit up to four differential diagnoses. Specialists and generalists achieved diagnostic accuracies of 38% and 19%, respectively, but both specialists and generalists were four percentage points less accurate for the diagnosis of images of dark skin as compared to light skin. Fair deep learning system decision support improved the diagnostic accuracy of both specialists and generalists by more than 33%, but exacerbated the gap in the diagnostic accuracy of generalists across skin tones. These results demonstrate that well-designed physician-machine partnerships can enhance the diagnostic accuracy of physicians, illustrating that success in improving overall diagnostic accuracy does not necessarily address bias.

Diagnostic precision and identification of rare diseases is dependent on distance of residence relative to tertiary medical facilities

BACKGROUND

Diagnostic precision and the identification of rare diseases is a daily challenge, which needs specialized expertise. We hypothesized, that there is a correlation between the distance of residence to the next tertiary medical facility with highly specialized care and the diagnostic precision, especially for rare diseases.

RESULTS

Using a nation-wide hospitalization database, we found a negative association between diagnostic diversity and travel time to the next tertiary referral hospital when including all cases throughout the overall International Classification of Diseases version 10 German Modification (ICD-10-GM) diagnosis codes. This was paralleled with a negative association of standardized incidence rates in all groups of rare diseases defined by the Orphanet rare disease nomenclature, except for rare teratologic and rare allergic diseases.

CONCLUSION

Our findings indicate a higher risk of being mis-, under- or late diagnosed especially in rare diseases when living more distant to a tertiary medical facility. Greater distance to the next tertiary medical facility basically increases the chance for hospitalization in a non-comprehensive regional hospital with less diagnostic capacity, and, thus, impacts on adapted health care access. Therefore, solutions for overcoming the distance to specialized care as an indicator of health care access are a major goal in the future.