January 8, 2026 By Kate Alfieri We’re thrilled to announce that Dr. Alexander Butwick has received a Department of Anesthesia Clinical Seed Award for the 2025-2026 cycle. The purpose of this award mechanism is to provide funding support to clinical Anesthesia faculty for research projects involving human subjects, human samples, and/or data derived from human subjects. This award is for faculty at all career stages and is suitable for individuals without substantial research experience. The Clinical Seed Award program is managed by the Anesthesia Biostatistics and Clinical Research Design (ABCD) Committee, which reviews applications and provides guidance and mentorship to applicants and awardees. The application is a multi-step process, involving a letter of intent, brief presentation, and full proposal submission over a 4–5-month period. During this process, applicants are given iterative feedback to help them develop the strongest project possible. Expert feedback continues during the award period, as awardees are required to give project updates throughout the year. More about Dr. Butwick’s project PI: Alexander Butwick, MBBS, MS ABCD Mentor: Romain Pirracchio, MD, PhD Study: Optimizing Postpartum Anemia Screening: Machine Learning Prediction of Hemoglobin Levels After Cesarean Delivery Postpartum anemia (PA) affects up to 50% of birthing patients in developed countries, leading to adverse outcomes including maternal depression, fatigue, impaired cognition, and difficulties with maternal-infant bonding and breastfeeding. Due to the lack of PA screening approaches, this study aims to develop the first prediction model for postpartum hemoglobin levels using ensemble machine learning techniques among US patients undergoing cesarean delivery (CD). The project has three main objectives: (1) To construct a comprehensive database of over 6,400 CD patients at UCSF (2016-2024), integrating detailed clinical and laboratory data; (2) To develop and validate an accurate machine learning model to predict postpartum hemoglobin levels after CD, using ensemble techniques such as random forests, gradient boosting, elastic net, and RuleFit; and (3) To qualitatively evaluate obstetricians' perspectives on a prediction model's accessibility, acceptability, and usability as a PA screening tool through semi-structured interviews. The long-term goal is to transform PA screening by integrating a validated prediction model into electronic health records across US maternity centers, enabling early identification and intervention for at-risk patients and reducing unnecessary laboratory testing for low-risk patients. The next Clinical Seed Award cycle is now open. Potential applicants are encouraged to reach out early by completing the ABCD intake form or emailing [email protected].
