New open-source dataset aims to help pharmaceutical companies provide faster, more accurate responses to healthcare provider inquiries about drugs
SAN FRANCISCO, March 22, 2024 /PRNewswire-PRWeb/ -- Ngram, a San Francisco-based startup specializing in generative AI solutions for the life sciences industry, today announced the release of its innovative dataset, medchat-qa, on Hugging Face. This dataset comprises a blend of real-world and synthetic questions healthcare providers (HCPs) frequently ask pharmaceutical companies about their drugs. It covers critical topics such as dosage, adverse reactions, drug interactions, new indications, and off-label uses. The release of medchat-qa marks a significant step in Ngram's mission to simplify the process for Medical Affairs teams to swiftly access literature and deliver quick, precise responses to HCPs.
Traditionally, doctors, nurses, pharmacists, and other HCPs seeking information about a drug must reach out to the pharmaceutical company's medical information department. Due to resource constraints, these inquiries can take days or even weeks to resolve, as the process involves navigating through extensive medical response documents and conducting labor-intensive literature reviews.
"HCPs require immediate access to accurate and current information to make informed treatment decisions," said Anish Muppalaneni, CEO and co-founder of Ngram. "However, the existing method for handling medical information requests is inefficient and slow. With medchat-qa, we aim to drive the development of AI systems capable of promptly highlighting pertinent information from scientific literature, enabling medical affairs teams to address HCP inquiries in seconds rather than days."
By making the medchat-qa dataset publicly available, Ngram intends to accelerate the development of question answering systems, dialogue agents, and other generative AI technologies for medical applications.
"Developing effective AI models necessitates access to substantial volumes of high-quality data that mirrors the real-world questions clinicians ask," explained Devadutta Ghat, CTO and co-founder of Ngram. "We have meticulously assembled medchat-qa to reflect the depth and complexity of medical information requests across various therapeutic areas. Releasing this dataset publicly allows us to leverage the collective expertise of the AI community to tackle this critical challenge."
The current process for responding to medical information inquiries is fraught with complexity and inefficiency. Medical Information professionals often rely on detailed MRDs to answer questions, resulting in the actual response being obscured within the document. This system leads to delays and dissatisfaction among HCPs and patients who receive generic responses that may not address their specific concerns.
"In today's fast-moving world, it's imperative that Medical Information be tailored for efficacy, and that data is optimized to minimize response times," Anish Muppalaneni added. "The medchat-qa dataset is a vital step toward fulfilling this objective and transforming how Medical Affairs teams deliver and access information for healthcare professionals."
Key Features of medchat-qa:
- Encompasses over 100,000 question-answer pairs.
- Covers more than 200 drugs across 100 therapeutic areas.
- Includes common queries such as dosage, drug interactions, contraindications, and adverse events.
- Serves as a benchmark dataset for assessing medical question answering system performance.
- Freely available on the Hugging Face dataset repository with MIT license..
The open-source release of the medchat-qa dataset on Hugging Face signifies a substantial progress in medical information access, highlighting Ngram's dedication to enhancing the healthcare experience for professionals and patients alike. The dataset can be accessed at this public url: https://huggingface.co/datasets/ngram/medchat-qa
To discover more about medchat-qa and Ngram's initiatives, visit www.ngram.com.
About Ngram
Ngram is dedicated to empowering life sciences organizations with generative AI technology. The company's solutions help Medical Affairs, Regulatory Affairs and Pharmacovigilance teams instantly access pertinent information, reducing response times from days to seconds. Founded in 2022 by Anish Muppalaneni and Devadutta Ghat, Ngram has received support from prominent investors in digital health and enterprise AI. For more information, visit www.ngram.com.
Media Contact:
Anish Muppalaneni
CEO, Ngram Inc
https://www.linkedin.com/in/anish-muppalaneni/
Email: [email protected]
Media Contact
Anish Muppalaneni, ngram Inc, [email protected], https://www.ngram.com
SOURCE Ngram
Share this article