About the company
Preverity, a healthcare data science company based in Nashville, Tennessee, provides advanced insights and risk data to its commercial clients that are reliant on quality data and risk assessments for physicians and other healthcare professionals. Preverity maintains one of the largest medical and pharmaceutical administrative transaction databases in the U.S, and leverages it for risk assessment through specialty-specific models in order to predict malpractice claims frequency for physicians and other professionals.
The challenge
Preverity’s vast database, containing over 6 billion rows of data, presents a significant challenge in terms of accessibility and usability for non-technical users. Hospital administrators and other stakeholders required a streamlined method to query this data for insights, particularly in identifying high-risk doctors, without needing to understand complex query languages or database structures. One of the large challenges here was using private LLM and appropriate infrastructure to maintain strict security protocols.
The process
Our team embarked on a collaborative project with Preverity to address this challenge by harnessing the power of modern Large Language Models (LLM). The goal was to create an intuitive interface that could translate natural language inquiries into structured SQL queries, and then interpret the database's responses back into natural, understandable language for the user. This process involved extensive training of the LLM to understand and accurately process medical and risk assessment-related queries, ensuring a high degree of precision and reliability in the responses generated.
Technologies
Large Language Models (LLM) for natural language understanding and generation
SQL for database querying
Custom integration tools to seamlessly connect the LLM with Preverity’s database infrastructure
Advanced security protocols to ensure data integrity and confidentiality of data
The solutions
We developed an advanced LLM-powered bot designed specifically for Preverity's needs. This bot allows users to input questions in natural language, which are then converted into SQL queries to retrieve data from Preverity’s extensive database. The system not only fetches the relevant data but also translates the complex results back into easily understandable natural language. This solution significantly reduces the barrier to accessing and interpreting large-scale healthcare data, enabling stakeholders to make informed decisions quickly and efficiently, while ensuring both security and performance.
Our impact
Our collaboration with Preverity has revolutionized the way healthcare risk data is accessed and interpreted, making it far more accessible to a broader range of users. This has led to:
01
Enhanced decision-making capabilities for hospital administrators and other stakeholders by providing immediate insights into physician risk profiles.
02
Increased efficiency in querying and interpreting vast amounts of data, saving valuable time and resources.
03
A scalable solution that can adapt to the growing size of Preverity’s database and the evolving needs of its users.
By leveraging the capabilities of LLM technologies, we have empowered Preverity to continue its mission of providing advanced risk assessment tools to the healthcare industry, thereby improving the quality and safety of healthcare delivery.