Purpose To conduct a prospective observational study across 12 U.S. hospitals to evaluate real-time performance of an interpretable artificial intelligence (AI) model to detect COVID-19 on chest radiographs. Materials and Methods A total of 95 363 chest radiographs were included in model training, external validation, and real-time validation. The model was deployed as a clinical decision support system, and performance was prospectively evaluated. There were 5335 total real-time predictions and a COVID-19 prevalence of 4.8% (258 of 5335). Model performance was assessed with use of receiver operating characteristic analysis, precision-recall curves, and F1 score. Logistic regression was used to evaluate the association of race and sex with AI model diagnostic accuracy. To compare model accuracy with the performance of board-certified radiologists, a third dataset of 1638 images was read independently by two radiologists. Results Participants positive for COVID-19 had higher COVID-19 diagnostic scores than participants negative for COVID-19 (median, 0.1 [IQR, 0.0–0.8] vs 0.0 [IQR, 0.0–0.1], respectively; P < .001). Real-time model performance was unchanged over 19 weeks of implementation (area under the receiver operating characteristic curve, 0.70; 95% CI: 0.66, 0.73). Model sensitivity was higher in men than women ( P = .01), whereas model specificity was higher in women ( P = .001). Sensitivity was higher for Asian ( P = .002) and Black ( P = .046) participants compared with White participants. The COVID-19 AI diagnostic system had worse accuracy (63.5% correct) compared with radiologist predictions (radiologist 1 = 67.8% correct, radiologist 2 = 68.6% correct; McNemar P < .001 for both). Conclusion AI-based tools have not yet reached full diagnostic potential for COVID-19 and underperform compared with radiologist prediction. Keywords: Diagnosis, Classification, Application Domain, Infection, Lung Supplemental material is available for this article. . © RSNA, 2022

Background Studies evaluating strategies for the rapid development, implementation, and evaluation of clinical decision support (CDS) systems supporting guidelines for diseases with a poor knowledge base, such as COVID-19, are limited. Objective We developed an anticoagulation clinical practice guideline (CPG) for COVID-19, which was delivered and scaled via CDS across a 12-hospital Midwest health care system. This study represents a preplanned 6-month postimplementation evaluation guided by the RE-AIM (Reach, Effectiveness, Adoption, Implementation, and Maintenance) framework. Methods The implementation outcomes evaluated were reach, adoption, implementation, and maintenance. To evaluate effectiveness, the association of CPG adherence on hospital admission with clinical outcomes was assessed via multivariable logistic regression and nearest neighbor propensity score matching. A time-to-event analysis was conducted. Sensitivity analyses were also conducted to evaluate the competing risk of death prior to intensive care unit (ICU) admission. The models were risk adjusted to account for age, gender, race/ethnicity, non-English speaking status, area deprivation index, month of admission, remdesivir treatment, tocilizumab treatment, steroid treatment, BMI, Elixhauser comorbidity index, oxygen saturation/fraction of inspired oxygen ratio, systolic blood pressure, respiratory rate, treating hospital, and source of admission. A preplanned subgroup analysis was also conducted in patients who had laboratory values (D-dimer, C-reactive protein, creatinine, and absolute neutrophil to absolute lymphocyte ratio) present. The primary effectiveness endpoint was the need for ICU admission within 48 hours of hospital admission. Results A total of 2503 patients were included in this study. CDS reach approached 95% during implementation. Adherence achieved a peak of 72% during implementation. Variation was noted in adoption across sites and nursing units. Adoption was the highest at hospitals that were specifically transformed to only provide care to patients with COVID-19 (COVID-19 cohorted hospitals; 74%-82%) and the lowest in academic settings (47%-55%). CPG delivery via the CDS system was associated with improved adherence (odds ratio [OR] 1.43, 95% CI 1.2-1.7; P<.001). Adherence with the anticoagulation CPG was associated with a significant reduction in the need for ICU admission within 48 hours (OR 0.39, 95% CI 0.30-0.51; P<.001) on multivariable logistic regression analysis. Similar findings were noted following 1:1 propensity score matching for patients who received adherent versus nonadherent care (21.5% vs 34.3% incidence of ICU admission within 48 hours; log-rank test P<.001). Conclusions Our institutional experience demonstrated that adherence with the institutional CPG delivered via the CDS system resulted in improved clinical outcomes for patients with COVID-19. CDS systems are an effective means to rapidly scale a CPG across a heterogeneous health care system. Further research is needed to investigate factors associated with adherence at low and high adopting sites and nursing units.

No abstract

Importance: An artificial intelligence (AI)-based model to predict COVID-19 likelihood from chest x-ray (CXR) findings can serve as an important adjunct to accelerate immediate clinical decision making and improve clinical decision making. Despite significant efforts, many limitations and biases exist in previously developed AI diagnostic models for COVID-19. Utilizing a large set of local and international CXR images, we developed an AI model with high performance on temporal and external validation. Objective: Investigate real-time performance of an AI-enabled COVID-19 diagnostic support system across a 12-hospital system. Design: Prospective observational study. Setting: Labeled frontal CXR images (samples of COVID-19 and non-COVID-19) from the M Health Fairview (Minnesota, USA), Valencian Region Medical ImageBank (Spain), MIMIC-CXR, Open-I 2013 Chest X-ray Collection, GitHub COVID-19 Image Data Collection (International), Indiana University (Indiana, USA), and Emory University (Georgia, USA) Participants: Internal (training, temporal, and real-time validation): 51,592 CXRs; Public: 27,424 CXRs; External (Indiana University): 10,002 CXRs; External (Emory University): 2002 CXRs Main Outcome and Measure: Model performance assessed via receiver operating characteristic (ROC), Precision-Recall curves, and F1 score. Results: Patients that were COVID-19 positive had significantly higher COVID-19 Diagnostic Scores (median .1 [IQR: 0.0-0.8] vs median 0.0 [IQR: 0.0-0.1], p < 0.001) than patients that were COVID-19 negative. Pre-implementation the AI-model performed well on temporal validation (AUROC 0.8) and external validation (AUROC 0.76 at Indiana U, AUROC 0.72 at Emory U). The model was noted to have unrealistic performance (AUROC > 0.95) using publicly available databases. Real-time model performance was unchanged over 19 weeks of implementation (AUROC 0.70). On subgroup analysis, the model had improved discrimination for patients with severe as compared to mild or moderate disease, p < 0.001. Model performance was highest in Asians and lowest in whites and similar between males and females. Conclusions and Relevance: AI-based diagnostic tools may serve as an adjunct, but not replacement, for clinical decision support of COVID-19 diagnosis, which largely hinges on exposure history, signs, and symptoms. While AI-based tools have not yet reached full diagnostic potential in COVID-19, they may still offer valuable information to clinicians taken into consideration along with clinical signs and symptoms.

No abstract