Treating complex patients poses significant challenges for clinicians, particularly when navigating nuanced clinical care scenarios. Boston Children’s Hospital addresses this challenge through the Clinical Doppelgangers project, which leverages large language models (LLMs) in conjunction with Amazon Web Services (AWS) to connect clinicians with patients sharing similar symptoms and conditions. This innovative approach enables healthcare providers to gain insights from past cases, enhancing their decision-making processes.

The project transforms unstructured clinical narratives into structured, queryable data, allowing clinicians to ask natural language questions and receive interpretable responses. By processing over 6,000 pediatric cases from the cardiac intensive care unit (CICU), the initiative has developed a comprehensive knowledge base incorporating more than 250 structured measurements and numerous semantic embeddings.

When clinicians seek specific patient cohorts, such as those with hypoplastic left heart and arrhythmia responsive to pacing, the LLM swiftly interprets their queries and executes a multi-step search. This expedites the identification of relevant patients, significantly enhancing decision-making for individual cases.

Utilizing AWS, the infrastructure supports efficient scaling and real-time processing. Key components include Amazon S3 for data storage, AWS Step Functions for workflow orchestration, and Amazon SageMaker for generating clinical embeddings. As a result, the Clinical Doppelgangers project is targeting an ambitious 80% reduction in the time required for chart review, thereby accelerating insights and improving clinical outcomes.

Looking ahead, the project aims to expand its applications beyond the CICU, integrate with electronic health records (EHR), and share its successful methodologies with other healthcare institutions, potentially transforming patient care in both pediatric and adult settings.

For centuries, pathology has revolved around the microscope, but a "historic transition" is now redefining the field, according to Dr. William Westra, vice chair of Anatomic Pathology at Moffitt Cancer Center. The shift from physical glass slides to digital imaging, bolstered by Artificial Intelligence (AI), is notably enhancing patient care and diagnostic efficiency.

This transformation is not merely technological; it reflects a change in professional identity. The traditional microscope has long been the cornerstone of pathology, but the introduction of digital scanning technology is reshaping observation methods. Pathologists now analyze high-resolution images on computer monitors rather than through microscope eyepieces, marking a significant leap in how tissue is examined.

The immediate benefits of digitization are profound. Westra cites enhanced accuracy and speed, noting that measuring the distance of a tumor from a margin, which once took minutes, can now be accomplished in mere seconds. Additionally, AI algorithms are integrated into the digital platform to identify cancer and assign objective tumor grades, greatly improving diagnostic efficiency.

Digital pathology also dissolves logistical challenges. Previously, acquiring second opinions or preparing for tumor board discussions involved cumbersome physical slide retrieval. Now, that process is instantaneous. Images can be shared globally within minutes rather than taking a week via traditional shipping methods, transforming interdepartmental operations.

Looking to the future, Westra believes digital pathology paired with deep learning will unlock new dimensions in oncology. AI's ability to identify critical biomarkers will enhance personalized medicine, moving beyond what is possible with traditional glass slides. Ultimately, this evolution will significantly improve patient care and usher in a new era of pathology.

As the Patient Monitoring series concludes, the dialogue transitions from current challenges to prospective innovations in healthcare. This fifth episode of the Health and Life Sciences at the Edge series envisions a future where effective patient monitoring transforms healthcare. Intel’s Kaeli Tully is joined by Sudha Yellapantula, Senior Researcher at Medical Informatics Corp., and Bikram Day, Director of Informatics at the same corporation, to discuss personalized care, bold predictions, and essential safeguards.

Yellapantula emphasizes real-world applications of AI, highlighting projects that utilize physiological data to predict cardiac arrest, evaluate CPR techniques, and identify critical events through signal analysis. She advocates for personalized predictive analytics tailored to individual data—encompassing history, genetics, and prior surgeries—to enhance patient outcomes.

Imagining healthcare in 2050, the episode envisions proactive, universal monitoring initiated upon hospital entry, where wearables can instantly flag major health events like strokes. This dynamic approach to monitoring aims to streamline alerts based on patient acuity, ensuring clinicians are only alerted to the most relevant information.

The conversation also predicts a shift in clinician workflows. Instead of navigating fragmented data, healthcare providers would access a comprehensive overview of each patient’s condition, eliminating logistical challenges. Day envisions the evolution of wearable technology that not only gathers data but also intelligently processes it, allowing for early intervention and equitable access to proactive care.

Central to this future is the importance of data integrity; Yellapantula warns that poor-quality training data can compromise AI effectiveness. The episode concludes with a powerful message: successful patient monitoring fosters early detection, informed decision-making, and individualized care, paving the way for improved patient outcomes through advanced technology and collaboration.

Hartford HealthCare has significantly enhanced access to primary care through its new virtual service, HHC 24/7, launched in partnership with K Health. Since its inception in April, HHC 24/7 has facilitated nearly 30,000 patient visits and garnered over 46,000 account registrations. The initiative addresses a critical demand for primary care services amid a statewide shortage, as patients often face long wait times for appointments. Dr. Padmanabhan Premkumar, president of Hartford HealthCare Medical Group, emphasizes the necessity of such services for timely patient care.

HHC 24/7 leverages AI technology to streamline patient interactions. When patients engage with the service, they begin by using an AI tool that generates their medical history for physicians and advanced practice providers. This innovation minimizes the hassle of repeating medical information, allowing for a more efficient consultation process. “Patients find it convenient, and we can see more patients per hour thanks to AI pre-visit charting,” says Premkumar, highlighting the pressure alleviation for providers.

Additionally, HHC 24/7 serves as an effective recruitment tool for clinicians who prefer working in a virtual environment. The current patient volume indicates a robust need for additional providers, further enhancing patient care access. Not only does the service provide immediate healthcare solutions, but it also connects patients to other Hartford HealthCare offerings, reinforcing a multi-faceted approach to healthcare delivery. Overall, HHC 24/7 is transforming patient and provider experiences, making high-quality care more accessible and efficient.