The AI Chronicles #5: GPTs, EU Clinical Trials Information System (CTIS), Med-PaLM2, and ClinicalGPT

The AI Chronicles 5 A Life Sciences - Pharma Edition
Reading Time: 4 mins

We are back with the 5th edition of The AI Chronicles Series. This blog discusses GPTs, the EU Clinical Trials Information System (CTIS), Med-PaLM2, ClinicalGPT, PMC-LLAMA, and MedAlpaca.

Do check out the previous four editions of the series:

The AI Chronicles #1 – What’s hot in the pharma world with Generative AI, how to build robust patient data privacy, and what’s up with Apple’s AI-powered health coach?
The AI Chronicles #2 – McKinsey’s 5-step digital transformation for Pharma companies, Sanofi’s AI approach to employee productivity, and more on digitization of medical data.
The AI Chronicles #3 – Find content in images and videos with Microsoft’s Vector, Washington’s new health privacy law, and the importance of personalization in healthcare.
The AI Chronicles #4 – AWS’s HealthScribe – a GenAI-powered service, a Roche-sponsored report, and Pytrial – a GenAI tool for various clinical trial tasks

GPTs!

OpenAI introduced GPTs, custom versions of ChatGPT designed for specific purposes. These versions, easily created without coding, will allow users to tailor ChatGPT to be more useful in their daily lives, work, or at home.

Close on the heels, Bill Gates wrote an insightful blog about how the advancements in Generative AI will completely change the way we use computers.

Gates highlights that current software requires users to specify which app to use for each task, but this will change in the next five years. He envisions a future where AI agents, responsive to natural language, will understand the user’s personal and professional life, eliminating the need to use different apps for different tasks.

These AI agents will be far more advanced than current technology, offering personalized assistance and capable of handling a variety of activities based on user permission. Gates believes these agents represent the biggest revolution in computing since the transition from typing commands to graphical interfaces.

Image source

What Does It Mean?

Combining this with OpenAI’s introduction of GPTs, a customizable version of ChatGPT for specific purposes, we see a trend towards more personalized, intelligent, and versatile AI tools. GPTs allow users to create tailored AI agents without coding knowledge, enabling various applications in daily life and work.

OpenAI’s focus on privacy and safety in GPT design complements Gates’ vision of AI agents deeply integrated into our personal and professional lives. Together, these developments suggest a future where AI agents, both GPTs and broader applications, will significantly change how we interact with technology, making it more intuitive, efficient, and personalized.

Transparency Rules & The EU Clinical Trials Information System (CTIS)

What is it about?

The European Medicines Agency (EMA) has initiated a public consultation to review the EU Clinical Trials Information System (CTIS) transparency rules. This review aims to balance the transparency of clinical trial information with confidentiality requirements while simplifying the use of CTIS to enhance user experience and minimize data breach risks.

The proposed changes, following public consultation, will uphold the transparency standards set by the Clinical Trials Regulation (CTR) and aim to reduce the administrative burden without compromising personal data and commercially confidential information protection. Additionally, the CTIS facilitates streamlined processes for applying for and supervising clinical trials across Europe, making it an attractive location for clinical research.

The system is a single entry point for clinical trial applications and assessments and includes a public database accessible to healthcare professionals and the public. The CTIS is maintained by the EMA, with the European Commission overseeing the implementation of the CT.

What does it mean?

Generative AI can greatly facilitate organizations’ adaptation to the updated EU Clinical Trials Information System (CTIS) rules by automating compliance checks, anonymizing sensitive data, streamlining reporting processes, and offering efficient training solutions. It enhances data analysis while ensuring adherence to transparency and confidentiality standards.

However, it’s crucial to employ advanced analytics techniques through a trusted partner to ensure data integrity and compliance with these evolving regulatory frameworks, thereby maintaining the delicate balance between transparency and confidentiality in clinical trials.

LLMs Everywhere In Healthcare

What is it about?

This paper provides a thorough and detailed survey emphasizing the transition from traditional Pretrained Language Models (PLMs) to more advanced LLMs. For the survey, the authors reviewed close to 18 models, including Med-PaLM2, ClinicalGPT, PMC-LLAMA, MedAlpaca, and more.

The survey highlights how LLMs enhance healthcare applications’ efficiency and effectiveness, comparing them with previous PLMs and among themselves. It discusses the advancements in LLMs, including their ability to address challenges previously faced by PLMs, such as lack of expressivity, difficulty in explanation, and the need for extensive data.

The Chain-of-Thought (CoT) ability and impressive few-shot or zero-shot capabilities of LLMs have improved their integration into healthcare.

What does it mean?

The article provides important directions for future development of generative AI applications in healthcare, including attempts at medical knowledge enhancement, including Retrieval-based LLMs, challenges with integrations, effective interaction with patients and doctors, and perhaps most importantly, dealing with hallucinations, misunderstandings, and prompt brittleness.

The Gramener Take

Gramener is an early GenAI adopter addressing various challenges across the life sciences/pharma/medical devices value chain including drug development and commercialization activities.

Some of our key solutions provide GenAI applications for Personal Identifiable Information (PII) anonymization, retrieval and summarizing of clinical development/medical affairs documents, KOL/HCP identification, engagement and management, promotional content generation, and pharma sales training.

Explore our tailored Generative AI solutions designed to do improved HCP engagements. Curious?

Leave a Comment

Your email address will not be published. Required fields are marked *

Scroll to Top
Share via
Copy link
Powered by Social Snap