ABSTRACT
Clinical trials are critical to many medical advances; however, recruiting patients remains a persistent obstacle. Automated clinical trial matching could expedite recruitment across all trial phases. We detail our initial efforts towards automating the matching process by linking realistic synthetic electronic health records to clinical trial eligibility criteria using natural language processing methods. We also demonstrate how the Sørensen-Dice Index can be adapted to quantify match quality between a patient and a clinical trial.
ABSTRACT
Purpose: Accurate classification of COVID-19 in chest radiographs is invaluable to hard-hit pandemic hot spots. Transfer learning techniques for images using well-known convolutional neural networks show promise in addressing this problem. These methods can significantly benefit from supplemental training on similar conditions, considering that there currently exists no widely available chest x-ray dataset on COVID-19. We evaluate whether targeted pretraining for similar tasks in radiography labeling improves classification performance in a sample radiograph dataset containing COVID-19 cases. Approach: We train a DenseNet121 to classify chest radiographs through six training schemes. Each training scheme is designed to incorporate cases from established datasets for general findings in chest radiography (CXR) and pneumonia, with a control scheme with no pretraining. The resulting six permutations are then trained and evaluated on a dataset of 1060 radiographs collected from 475 patients after March 2020, containing 801 images of laboratory-confirmed COVID-19 cases. Results: Sequential training phases yielded substantial improvement in classification accuracy compared to a baseline of standard transfer learning with ImageNet parameters. The test set area under the receiver operating characteristic curve for COVID-19 classification improved from 0.757 in the control to 0.857 for the optimal training scheme in the available images. Conclusions: We achieve COVID-19 classification accuracies comparable to previous benchmarks of pneumonia classification. Deliberate sequential training, rather than pooling datasets, is critical in training effective COVID-19 classifiers within the limitations of early datasets. These findings bring clinical-grade classification through CXR within reach for more regions impacted by COVID-19.
ABSTRACT
Outpatient centers comprised of many concurrent clinics increasingly see higher patient volumes. In these centers, decisions to improve clinic flow must account for the high degree of interdependence when critical personnel or equipment is shared between clinics. Discrete event simulation models have provided clinical decision support, but rarely address high-volume clinics with shared resources. While highly complex models are now capable of representing clinics in detail, validation techniques often do not evaluate model predictive performance when presented with new data. Cross-validation provides a means to evaluate the robustness of model treatment time predictions when ongoing data collection in clinics is impractical. Ensuring robust predictions assures validity in the use of models to optimize clinic performance. We apply cross-validation in evaluating a model of glaucoma clinic service at Duke Eye Center. In-person observation is used to verify the accuracy of operations data collected through electronic health records (EHR). From the EHR data, we formulate a stochastic reward net model, employing phase-type distributions to represent treatment durations, and solved through discrete event simulation. The model is formulated in two configurations to represent (1) concurrent demand on clinic staff, or (2) independently functioning clinics. Evaluating these two alternatives in cross-validation studies, we find model prediction accuracy improves when interdependence is explicitly modeled in the examined setting.
