We are no longer accepting abstracts for SAIL 2022.
The Symposium on Artificial Intelligence for Learning Health Systems (SAIL) is an annual international conference exploring the integration of artificial intelligence (AI) techniques into clinical medicine. SAIL will be held in Hamilton, Bermuda on May 22–24, 2022.
An explicit goal of this conference is to foster collaboration between methodologists and key healthcare decision makers such as clinicians, hospital administrators, regulators and payors. SAIL will feature leading voices from each of these backgrounds to shed light on opportunities for and obstacles to the deployment of clinical AI systems.
A second key goal of SAIL is to better integrate the complementary experience of the clinical informatics and machine learning communities. Clinical informaticians, for their part, have an extensive history of integrating their tools within hospital IT systems and working with administrators, privacy policies, and quality improvement officers. Computer scientists, for their part, have frequently spearheaded the development of novel methodological tools and high-performance computing. SAIL is designed to bring together the best of each of these closely related but often disconnected communities.
SAIL will feature invited presentations to expose AI practitioners to the clinical workflow and administrative challenges that commonly prevent real-world adoption. Panels will convene seasoned leaders who have overseen the implementation, adoption, and regulation of real clinical AI systems in practice. Tutorials will provide hands-on exposure to open-source tools for integrating apps with hospital IT systems. Finally, we solicited abstracts for podium or poster presentations designed to generate fruitful discussion (and debate!) among conference attendees from diverse backgrounds (clinicians, clinical informaticians, computer scientists, payers, and regulators).
Abstracts contained either: 1) new and unpublished work, 2) highlights of recently published work or 3) important research hypotheses or questions. Work describing pure methods without thoughtful consideration of clinical applicability or implementation were considered out of scope. The ideal submission used innovative machine learning methodology to address real world clinical problems with results that can be translated to improving healthcare. Abstracts needed to be well supported with references to prior work.
Example research themes included addressing key challenges in combining AI with clinical workflows (e.g., clinical decision support, human computer interaction, interpretability, active learning, etc.) or deploying intelligent systems at scale (e.g., fairness, privacy, generalizability across institutions over time, real-time prediction, regulatory compliance). We highly encouraged work that involves interdisciplinary collaboration across AI researchers, clinicians, and informaticians. The following are examples of highly rated submissions from the 2020 virtual presymposium:
- Laura Early Warning System (LEWS) - An Artificial Intelligence (AI) Platform for the Management of Clinical Deterioration on the Wards - Preliminary Results from Brazilian Hospitals
- System for High Intensity EvaLuation During Radiation Therapy (SHIELD-RT): A prospective randomized study of machine learning-directed clinical evaluations during outpatient cancer radiation and chemoradiation
- Standard ML Models on MIMIC Fail to Generalize over Time
Abstracts had a 500-word limit, excluding references, figures, and figure captions. Abstract submissions were unblinded and could be submitted in any format, and could contain content that had been published within the last 12 months. Abstracts were selected for poster and oral presentations by the Program Committee. In addition to being invited to present their work in-person at the symposium, all accepted submissions will also be posted on the SAIL website. View the 2020 Virtual Presymposium posters and presentations.