Search

PurposeTo develop and evaluate an automated, portable algorithm to differentiate active corneal ulcers from healed scars using only external photographs.DesignA convolutional neural network was trained and tested using photographs of corneal ulcers and scars.ParticipantsDe-identified photographs of corneal ulcers were obtained from the Steroids for Corneal Ulcers Trial (SCUT), Mycotic Ulcer Treatment Trial (MUTT), and Byers Eye Institute at Stanford University.MethodsPhotographs of corneal ulcers (n = 1313) and scars (n = 1132) from the SCUT and MUTT were used to train a convolutional neural network (CNN). The CNN was tested on 2 different patient populations from eye clinics in India (n = 200) and the Byers Eye Institute at Stanford University (n = 101). Accuracy was evaluated against gold standard clinical classifications. Feature importances for the trained model were visualized using gradient-weighted class activation mapping.Main outcome measuresAccuracy of the CNN was assessed via F1 score. The area under the receiver operating characteristic (ROC) curve (AUC) was used to measure the precision-recall trade-off.ResultsThe CNN correctly classified 115 of 123 active ulcers and 65 of 77 scars in patients with corneal ulcer from India (F1 score, 92.0% [95% confidence interval (CI), 88.2%-95.8%]; sensitivity, 93.5% [95% CI, 89.1%-97.9%]; specificity, 84.42% [95% CI, 79.42%-89.42%]; ROC: AUC, 0.9731). The CNN correctly classified 43 of 55 active ulcers and 42 of 46 scars in patients with corneal ulcers from Northern California (F1 score, 84.3% [95% CI, 77.2%-91.4%]; sensitivity, 78.2% [95% CI, 67.3%-89.1%]; specificity, 91.3% [95% CI, 85.8%-96.8%]; ROC: AUC, 0.9474). The CNN visualizations correlated with clinically relevant features such as corneal infiltrate, hypopyon, and conjunctival injection.ConclusionsThe CNN classified corneal ulcers and scars with high accuracy and generalized to patient populations outside of its training data. The CNN focused on clinically relevant features when it made a diagnosis. The CNN demonstrated potential as an inexpensive diagnostic approach that may aid triage in communities with limited access to eye care.

To compare disparities in outpatient ophthalmic care during early and later periods of the COVID-19 public health emergency. This cross-sectional study compared non-peri-operative outpatient ophthalmology visits by unique patients at an adult ophthalmology practice affiliated with a tertiary-care academic medical center in the Western US during three time periods: pre-COVID (3/15/19-4/15/19), early-COVID (3/15/20-4/15/20), and late-COVID (3/15/21-4/15/21). Differences in participant demographics, barriers to care, visit modality (telehealth, in person), and subspeciality of care were studied using unadjusted and adjusted models. There were 3095, 1172 and 3338 unique patient-visits during pre-COVID, early-COVID and late-COVID (overall age 59.5 ± 20.5 years, 57% female, 41.8% White, 25.9% Asian, 16.1% Hispanic). There were disparities in patient age (55.4 ± 21.8 vs. 60.2 ± 19.9 years), race (21.9% vs. 26.9% Asian), ethnicity (18.3% Hispanic vs. 15.2% Hispanic), and insurance (35.9% vs. 45.1% Medicare) as well as changes in modality (14.2% vs. 0% telehealth) and subspecialty (61.6% vs. 70.1% internal exam specialty) in early-COVID vs. pre-COVID (p <.05 for all). In late-COVID, only insurance (42.7% vs. 45.1% Medicare) and modality of care (1.8% vs. 0% telehealth) persisted as differences compared to pre-COVID. There were disparities in patients receiving outpatient ophthalmology care during early-COVID that returned close to pre-COVID baseline one year later. These results suggest that there has not been a lasting positive or negative disruptive effect of the COVID-19 pandemic on disparities in outpatient ophthalmic care. [ABSTRACT FROM AUTHOR]

Background: Factors impeding delivery of adequate enteral nutrition (EN) to trauma patients include delayed EN initiation, frequent surgeries and procedures, and postoperative ileus. We employed 3 feeding strategies to optimize EN delivery: (1) early EN initiation, (2) preoperative no nil per os feeding protocol, and (3) a catch-up feeding protocol. This study compared nutrition adequacy and clinical outcomes before and after implementation of these feeding strategies., Methods: All trauma patients aged ≥18 years requiring mechanical ventilation for ≥7 days and receiving EN were included. Patients who sustained nonsurvivable injuries, received parenteral nutrition, or were readmitted to the intensive care unit (ICU) were excluded. EN data were collected until patients received an oral diet or were discharged from the ICU. The improvement was quantified by comparing nutrition adequacy and outcomes between April 2014-May 2015 (intervention) and May 2012-June 2013 (baseline)., Results: The intervention group (n = 118) received significantly more calories (94% vs 75%, P < .001) and protein (104% vs 74%, P < .001) than the baseline group (n = 121). The percentage of patients receiving EN within 24 and 48 hours of ICU admission increased from 41% to 70% and from 79% to 96% respectively after intervention (P < .001). Although there were fewer 28-ay ventilator-free days in the intervention group than in the baseline group (12 vs 16 days, P = .03), receipt of the intervention was associated with a significant reduction in pneumonia (odds ratio, 0.53; 95% confidence interval, 0.31-0.89; P = .017) after adjusting sex and Injury Severity Score., Conclusions: Implementation of multitargeted feeding strategies resulted in a significant increase in nutrition adequacy and a significant reduction in pneumonia., (© 2017 American Society for Parenteral and Enteral Nutrition.)