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.

Identification of Racial Inequities in Access to Specialized Inpatient Heart Failure Care at an Academic Medical Center.

BACKGROUND: Racial inequities for patients with heart failure (HF) have been widely documented. HF patients who receive cardiology care during a hospital admission have better outcomes. It is unknown whether there are differences in admission to a cardiology or general medicine service by race. This study examined the relationship between race and admission service, and its effect on 30-day readmission and mortality Methods: We performed a retrospective cohort study from September 2008 to November 2017 at a single large urban academic referral center of all patients self-referred to the emergency department and admitted to either the cardiology or general medicine service with a principal diagnosis of HF, who self-identified as white, black, or Latinx. We used multivariable generalized estimating equation models to assess the relationship between race and admission to the cardiology service. We used Cox regression to assess the association between race, admission service, and 30-day readmission and mortality. RESULTS: Among 1967 unique patients (66.7% white, 23.6% black, and 9.7% Latinx), black and Latinx patients had lower rates of admission to the cardiology service than white patients (adjusted rate ratio, 0.91; 95% CI, 0.84-0.98, for black; adjusted rate ratio, 0.83; 95% CI, 0.72-0.97 for Latinx). Female sex and age >75 years were also independently associated with lower rates of admission to the cardiology service. Admission to the cardiology service was independently associated with decreased readmission within 30 days, independent of race. CONCLUSIONS: Black and Latinx patients were less likely to be admitted to cardiology for HF care. This inequity may, in part, drive racial inequities in HF outcomes.

Successful Use of Cyclosporin A for Stevens-Johnson Syndrome and Toxic Epidermal Necrolysis in Three Children.

BACKGROUND/OBJECTIVES: Stevens-Johnson syndrome (SJS) and toxic epidermal necrolysis (TEN) are medical emergencies. Mainstays of treatment include removal of the offending agent, supportive care, and wound care. The use of immunosuppressive agents such as corticosteroids and intravenous immunoglobulin (IVIg) is controversial. Some case reports and small studies report the successful use of cyclosporin A (CsA) for SJS/TEN in halting disease progression, fostering reepithelialization, and reducing mortality. OBJECTIVE: To report on the efficacy of cyclosporine A in the treatment of SJS/TEN in three pediatric patients. METHODS: We describe three pediatric patients seen at a tertiary care hospital in Boston, Massachusetts, diagnosed with SJS/TEN confirmed by skin biopsy who were successfully treated with CsA with improvements seen in time to cessation of disease progression or new lesion formation, reepithelialization, and duration of hospital stay. RESULTS: The average time cessation of disease progression or new lesion formation after CsA administration was 2.2 days (range 1.5-3 days) and average time to remission or reepithelialization was 13 days (range 10-15 days). The average length of hospital stay was 11.7 days (range 4-19 days). CONCLUSIONS: We describe three pediatric patients treated successfully with CsA and provide evidence for the use of cyclosporine in children with SJS/TEN. These results further support previous observations that CsA use for SJS/TEN produces consistently favorable outcomes. The results in this case series are limited by their observational nature. Additional trials are needed to evaluate the safety and efficacy of CsA use in children.

Radiographic alignment of unstable distal radius fractures fixed with 1 or 2 rows of screws in volar locking plates.

PURPOSE: We tested the null hypothesis that there is no difference in the change in volar tilt of the articular surface of the distal radius on lateral radiographs obtained before suture removal and 3 months or more after surgical fracture fixation when 1 or 2 rows of screws are used in the distal part of a volar locked plate. METHODS: We retrospectively identified 364 consecutive patients with a distal radius fracture treated by 2 surgeons with open reduction and volar locked plate fixation between 2007 and 2011 at our institution. A manually case-matched design with one-to-one matching of 2 different strategies for screws in the distal part of the plate (1 row versus 2 rows) on the basis of sex, AO type, presence of dorsal comminution, ulna fracture, mechanism of injury, and age (± 8 y) resulted in a group of 34 pairs, 68 total fractures. Radiographic alignment was measured before suture removal and 3 months or more after surgery. RESULTS: The change in volar tilt of the articular surface was -1.2° in the 1-row group and -0.9° in the 2-row group, which was not significantly different. The secondary displacement of radial inclination and ulnar variance were likewise small and not statistically significant. CONCLUSIONS: We found no advantage of 2 rows of distal screws over a single row of screws with respect to maintenance of achieved restoration of volar angulation after volar locked plate fixation of AO type A and C fractures of the distal radius. TYPE OF STUDY/LEVEL OF EVIDENCE: Therapeutic III.