Transforming hearing care with machine learning

The Challenge: Proactive risk assessment in hearing health

A prominent US-based hearing benefit provider sought to enhance its ability to proactively identify members at high risk of Sensorineural Hearing Loss (SNHL), particularly those over 40.  Their existing methods presented several key challenges:

• Lack of proactive identification & delayed intervention: Existing reactive methods led to delayed diagnoses and missed opportunities for early, proactive intervention.
  
  
• Operational inefficiencies: This reactive approach resulted in less efficient care delivery.
  
  
• Incomplete member health insights: There was a less comprehensive understanding of overall member health.
  
  
• Challenges in engagement & partnerships: These limitations hindered their ability to effectively engage members and form strategic partnerships.

The primary goal was to establish a system that could accurately predict risk and facilitate timely audiology screenings.

The Solution: An intelligent, AI-powered platform

To address these challenges, CitiusTech collaborated with the hearing benefit provider to design and implement a sophisticated, intelligent platform leveraging AI and machine learning (ML).

Exhibit 1: Solution Highlight

Core solution highlights

• Predictive modeling

• A robust ML model was developed to predict SNHL Hearing Loss Risk Scores and categorize risk for members above the age of 40 years who did not yet have a hearing loss diagnosis.
  
  
• Key factors contributing to hearing loss were identified and integrated into the model, enhancing its predictive accuracy.
  
  

• ML Ops pipeline for automation

• An end-to-end (E2E) CI/CD automation pipeline was established using AWS Sagemaker projects and pipelines. This ensured seamless model deployment, continuous performance, scalability, and included lineage tracking and auditing capabilities.
  
  

• Analytics dashboard with model monitoring

• An interactive analytics dashboard was created with AWS QuickSight, providing real-time visibility into the distribution of members with and without hearing loss across various recognized aspects.
  
  
• The dashboard also showcased crucial model metrics and infrastructure metrics, offering comprehensive performance oversight.
  
  

• Claim dollar utilization and risk assessment

• The platform computed the "all claims" PMPM (Per Member Per Month) utilization.
  
  
• It also analyzed hearing loss-associated co-morbid conditions such as fall, depression, and dementia, and chronic diseases like Diabetes and Hypertension.
  
  
• Furthermore, the system computed the Risk Adjustment Factor (RAF) to demonstrate the recent health risk of members and a co-morbid severity score specifically for members with SNHL.

The Impact: Superior care, strategic growth, and enhanced efficiency

The implementation of the AI-powered platform had a significant impact on the hearing benefit provider's operations and member care:

• Improved member identification for audiology screening: The ML model accurately identified high-risk members, enabling proactive outreach and timely audiology screenings. This shift from reactive to preventive care directly improved the quality of care and facilitated earlier intervention.
  
  
• Enhanced understanding of member risk: The platform provided a comprehensive view of member health, integrating data on claim utilization, Risk Adjustment Factor (RAF), and co-morbidity severity scores. This holistic insight supported more informed decision-making in care management.
  
  
• Optimized care pathways: With AI-driven insights, the provider could streamline and optimize care interventions, ensuring resources were allocated effectively to those most in need.
  
  
• Strengthened Payer relationships:The demonstrable success in identifying at-risk members and improving outcomes provided a strong value proposition to payers, leading to increased confidence and fostering new partnerships.

What does this mean for other healthcare organizations?

This engagement demonstrates how a focused application of AI can directly address core challenges within healthcare operations. By leveraging advanced machine learning, organizations can:

• Proactively identify and manage patient risk: Shifting from reactive care to predictive interventions.
  
  
• Gain deeper insights from complex data: Unlocking actionable intelligence from clinical and operational data.
  
  
• Optimize resource allocation and care delivery: Ensuring efficiency and effectiveness in patient management.
  
  
• Strengthen relationships with stakeholders: Demonstrating value through improved outcomes and operational excellence.