Search for: All records

Sepsis is a dysregulated host response to infection with high mortality and morbidity. Early detection and intervention have been shown to improve patient outcomes, but existing computational models relying on structured electronic health record data often miss contextual information from unstructured clinical notes. This study introduces COMPOSER-LLM, an open-source large language model (LLM) integrated with the COMPOSER model to enhance early sepsis prediction. For high-uncertainty predictions, the LLM extracts additional context to assess sepsis-mimics, improving accuracy. Evaluated on 2500 patient encounters, COMPOSER-LLM achieved a sensitivity of 72.1%, positive predictive value of 52.9%, F-1 score of 61.0%, and 0.0087 false alarms per patient hour, outperforming the standalone COMPOSER model. Prospective validation yielded similar results. Manual chart review found 62% of false positives had bacterial infections, demonstrating potential clinical utility. Our findings suggest that integrating LLMs with traditional models can enhance predictive performance by leveraging unstructured data, representing a significant advance in healthcare analytics. 

We demonstrate a compact multilayer GaAs–AlAs structure for passive optical edge detection at multiple wavelengths. Through the inverse design of the layer thicknesses, this structure manipulates spatial frequency components of an incoming wavefront, selectively reflecting high-frequency features while suppressing low-frequency intensity variations. Simulations reveal a reflectance transition from minimal to near-total as a function of numerical aperture, a property leveraged for enhancing edge contrast in optical imaging. For the first time, to our knowledge, we utilize molecular beam epitaxy (MBE) to fabricate edge detection devices, ensuring structural fidelity. Material characterization confirms high-quality interfaces, precise thickness control, and excellent uniformity, validating the suitability of MBE for this application. Experimental angle-resolved reflectance measurements closely align with theoretical predictions, demonstrating the feasibility of this approach for real-time, hardware-based optical image processing. The proposed design automatically works for at least two wavelengths and can be readily extended to operate at multiple wavelengths simultaneously. This work opens new possibilities for employing multilayer interference structures in high-performance optical imaging and real-time signal processing.