Problem

In oncology (and healthcare in general), an overwhelming amount of data is generated during diagnosis, treatment and monitoring of patients and this vast amount of data comes from various sources including EHRs, imaging, genomics sequencing and more.

The different sources of data lead to variations in format, quality and reliability making it challenging to integrate and compare accurately. The complexity and sheer quantity creates challenges for those working with this data.

From a clinical perspective, it is hard to extract insights from the data and therefore difficult to make timely and informed decisions. For those using the data for operational purposes, incomplete data leads to inaccuracies in reporting.

Solution

As the UX lead for Analytics, my core responsibility for this project was to redesign the existing interface in order to enhance and improve our core user needs:

• Allow data customization: This is a highly requested feature from our customers. Operationally, if we can enable customers to create their own dashboards, reports and exports using Navify Analytics, they will be able to do all their analysis within the product and they will pay for the value this provides.
   
• Improve access to clinical insights: If we enable clinicians to glean data-driven insights either by asking key questions or following a train of thought, they will value and use those insights for treatment recommendations and pay for the value this provides.
   
• Leverage AI/NLP: Use Gen AI solutions or any other LLM models to help optimize analytic workflows. Based on the pattern of usage, the system may suggest relevant downstream actions. (For example, if the Clinical coordinator prepares a file for accreditation export, the system can suggest options on how to send that data directly to the CoC for accreditation.)

---

Personas

There are multiple personas in a healthcare setting that have a use for the oncology data collected by our products. Certified Tumor Registrars (CTRs) analyze the data to adhere to regulations and accreditations and report both externally and internally.

Currently, CTRs and Clinical Coordinators are the most common users of Navify Analytics tools and data in Navify Clinical Hub.

Carol, the Cancer Tumor Registrar (CTR)

---

Voice of the customer

• In order to understand our core user segment, we set up a 1-day Tumor registrar workshop at Roche to learn about the various day-to-day tasks and pain points from 6 different current registrars across the U.S.
   
• The goal was to collect input from and validate unknowns with customers (or customer representatives) to ensure we are creating value for those who use our products.
   
• I partnered with the CX team to develop a journey map to help visualize a deeper understanding of the user’s emotions, motivations, and pain points throughout their interaction with the product.

---

Final redesign

The redesign of Navify Analytics will improve user satisfaction, better decision-making, and ultimately, enhanced patient care outcomes.

Enhanced patient list

Ensure the data fields needed for an effective search, but limit the fields only to identifying the patient

• Improved search: Advanced search UI options
   
• Enhanced filters: Refine the searchable records, allow saving of filter sets
   
• Feature flag capabilities: Display data elements specific to Tumor Board or Oncology Navigator bundles. (Add new columns as needed)

Aggregated Tumor Board data

Navify Analytics allows access to all relevant patient clinical insight data from both Oncology Navigator and integrated Tumor Boards.

• One patient list: Combine all data in one simplified display format. Allow users to access both current patient data with tumor board data.
   
• Advanced tumor board reporting: Integrate a new category for Tumor Board reports.

Powerful dashboards

Analytics dashboards help streamline analysis and discover key insights in an easy to scan visual data format.

• Enhanced data visualizations: transforms complex datasets into easily understandable visual formats, making it easier for users to grasp insights quickly.
   
• Improved decision making: Enable faster comprehension of critical data, facilitating quicker decision-making and response times.

Enhanced usability

• Simplify the navigation and ensure the UI is easy to understand for users with varying levels of technical expertise.
   
• Present data in a clear, accurate, and meaningful way to facilitate quick insights and decision-making

Leverage/share existing technology where possible

• Consider consistency and best practices across the entire Navify portfolio – explore opportunities to re-use technology
   
• Create best practices/standards for Navify Analytics, and the entire Navify portfolio
   
• Future flexibility: Consider that Analytics will expand beyond the Oncology disease area network

---

AI/NLP concept explorations

Chatbots and virtual agents

Explore opportunities in Navify Analytics for a conversational chatbot or a smart search engine that could be NLP-based or further expanded to use Gen AI. It can work in the following way:

• Users will give their intended input through text or voice in simple English on what patient or meeting cohort they intend to see. This can be a combination of filters, column customizations or even exports.
   
• The pre-trained model will understand the user’s intent and give out the result accordingly. 
   
• The user can open the filter pop-up to validate their request if needed – the appropriate filters will be checked.

This can be further expanded with Gen AI which apart from the user input system will give out suggestions for better results.

For example, a user is trying to search for patients in a 6-month time range with a certain set of primary cancers and their associated biomarkers. This search turns out to be null. The scope of an NLP will end here.

However, with Gen AI we can give suggestions to users on filter sets where they can get better results closer to their intent.  If they do get a result back from their search, they could ask GenAI to suggest other data they should be looking at or point out unique data points in their list of patients.

Additional AI opportunities:

• Auto-generation of Visualizations and Dashboard Layouts: Using Gen AI the system should suggest and represent the rightful charts for the respective data selected.
  
• Automation and Workflows: We can use Gen AI solutions where we can train GPT or any other LLM models to adopt the persona of a customer service agent who uses the Analytics for their daily purposes. 

Note: This project is currently ongoing and some screens are not available for public view.
Contact Studionaka for the full case study