From Data to Decisions

Where we are in the insights journey for healthcare

Healthcare industry has generated massive data over the last few decades, and it will continue to exponentially increase the variety and volume of the data being generated. Several healthcare organizations took up this challenge and implemented data platforms that bring the data together on modern data platforms like Snowflake; many have even gone further to create data assets and products that are supposedly aligned to the business needs.

However, deriving business insights from all the data and making it accessible to the wide business community to enable data driven decision making, continues to be a challenge – mainly due to the following reasons:

• Expensive methods of delivery - traditional methods of delivering such actionable insights have been through reports and dashboards built by armies of data and BI analysts.
• Continuously evolving – business needs keep evolving and demand for business insights continuously increases, the speed to building reports and dashboards can’t keep up with the business
• Disconnected dots – a lot of valuable insight is locked up in documents, images and other unstructured data and a full picture is often missing.
• Rapid iteration – a new business insight frequently leads to a need for the next level of insights or identifies a need for completely different insight to make a business decision or take an action, the delivery of reports or even query results can’t match the speed the business needs them at.

Across healthcare, there are several scenarios that could be impacted by lack of rapid, accurate and iterative insights; some of the examples are:

• Care gaps identification and managing cohorts by taking targeted action for complex population health challenges like diabetes requires iterative insights and uncovering data from patient charts. If left unaddressed, care gaps could result in deterioration of health-plan quality scores, significant reductions in quality incentive payments and even adverse impact to patients’ health.
• Design of alternative payment models for value-based care requires comparison of provider performances over time across different patient conditions and patterns. Many insights needed may not be defined clearly upfront and require several iterations for each business scenarios based on constraints embedded deep in policy documentation.
• Fraud Waste & Abuse Analysis - Identification of potential fraud and waste through upcoding might require analyses of claims data and ICD’s against clinical notes in patient charts. Integrating the unstructured data to claims would typically require data science skills and scaling it, for possibly the entire population of lives covered, could take a massive effort.
• Practice Profitability: Analysis of profitability by procedure, patient, or physician, as well as optimizing operational efficiency in areas like the operating room and surgical scheduling. Empower physicians to understand financial metrics through intelligent Dashboards.
• Quality Management & Care Variation: Benchmark physician and facility performance and identify high-quality care providers, helping to reduce variations in care and manage quality. AI driven dashboards enable physicians to improve treatment plans and outcomes.

Snowflake Intelligence as a solution to these challenges

Snowflake Intelligence is an agentic experience accessible via a dedicated portal that addresses the core challenges – i.e.

• Uncovering truly comprehensive 360-degree insights by combining unstructured data with the already known data points
• Making insights accessible to anyone who needs them through a conversational interface
• Delivering insights instantly without relying on a team to write queries or build a report

Snowflake Intelligence is implemented as an ‘Agent’ that leverages the powerful capabilities of Cortex AI; specifically, Cortex Analyst and Cortex search to generate the insights.

The Snowflake Intelligence agent takes a natural language question and identifies the best course of action to arrive at a requested insight, manages routing to leverage structured or unstructured data or both as needed, and finally provides a business usable response, including any relevant visualizations. Intelligence relies on a semantic model to instruct the agent on relevant business semantics, build relevant context and provide accurate responses.

Snowflake Intelligence also provides the flexibility to pick an underlying foundational GenAI model to suit the preferences of an organization and at the same time maintains the robust access controls on data.

CitiusTech’s core focus on Healthcare with Snowflake Intelligence and Cortex services is a perfect alliance to address and replace expensive BI solutions. CitiusTech has launched “Healthcare Insights” agents using Snowflake Intelligence platform and Cortex to help healthcare organizations to rapidly iterate on data insights, making them actionable and truly useful for their patients care journey & outcomes.

Through “Healthcare Insights” agents we are positively impacting a leading Blues’ Star ratings campaign. Star campaign managers are now able to work on incoming Insights asks from care managers for specific measures in less than an hour which normally took him 2-3 days through the traditional BI dashboards.

Key Highlights of CitiusTech solution

1. re-usable and deployable use-case specific semantic models and agents
2. Integrated Intelligence with an award-winning agentic guardrail “Quality & Trust”
3. Natively built using Snowflake Intelligence and Cortex