Streamlining Prior Authorization with Treatline’s Generative AI Platform for Healthcare and Insurance Providers
By Artem Kobrin, Head of Cloud Practice & Partner – Neurons Lab
By Dima Solopov, Head of AI Solutions Practice & Partner – Neurons Lab
By Qiong Zhang, Sr. Partner Solutions Architect – AWS
Prior authorization is an insurance review process used by healthcare providers to confirm medical coverage with health insurance companies.
This process involves submitting detailed documentation and clinical justifications to insurance companies, which puts significant strain on healthcare providers. This can result in care delays, negative clinical outcomes, and increased costs, with the administrative cost for physicians processing prior authorization requests amounting to $528 million in 2019.
In this post, we will outline how Neurons Lab, an AWS Advanced Tier Services Partner and AWS Marketplace Seller with Competencies in Machine Learning and Healthcare, collaborated with Treatline on its platform to streamline the highly manual prior authorization process using intelligent document processing (IDP) and generative AI.
The Treatline platform facilitates the exchange of prior authorization requests through automation and improved payer/provider communication. It also provides analytical tools to measure treatment, clinician, facility, and organizational-level performance and comply with recent changes in legislation.
Treatline’s mission is to reduce the burden of prior authorization on healthcare providers and insurance companies, ensuring patients receive timely and appropriate care while streamlining the overall process.
With a strong background in both healthcare and technology, the Treatline team is uniquely positioned to address the challenges faced by medical staff, patients, and insurance companies alike.
By focusing on improving patient outcomes, reducing administrative burden, and streamlining the healthcare delivery process, Treatline is committed to making a significant impact on the healthcare and insurance industries. Its innovative platform bridges the gap between healthcare providers and insurance companies, promoting greater efficiency and collaboration to ensure the best possible outcomes for all parties involved.
Challenge: Inefficiencies in Prior Authorization
Prior authorization is a pre-approval process required by insurance companies to confirm the medical necessity and coverage of specific treatments, procedures, medications, or services before they are performed or provided. It involves submitting detailed documentation and clinical justifications to the insurance company, demonstrating the need for the proposed healthcare intervention.
Prior authorization is often a time-consuming process that leads to staff burnout and negative impacts on patient care. Physicians spend an average of 14 hours per week on prior authorization tasks, submitting around 40 requests per week. Additionally, 90% of physicians describe the burden associated with prior authorization as high or extremely high.
Patients also suffer, with 93% of physicians reporting care delays, and 42% stating that prior authorization has led to serious adverse events for their patients, such as hospitalization, life-threatening events, or even death.
Insurance companies are affected by the inefficiencies in the prior authorization process as well. Delays in claim approvals and the complex, manual nature of the process contribute to increased operational costs and dissatisfied customers.
Despite 90-95% of prior authorization requests being approved eventually, the process involves time-consuming steps for both healthcare providers and insurance companies, including collecting data, submitting requests, and handling denials and appeals. These inefficiencies call for a streamlined solution that benefits all parties involved.
Solution: Automation and AI-Powered Platform
Treatline’s platform, developed in partnership with Neurons Lab, leverages cutting-edge AI technology and AWS services to streamline the prior authorization process for insurance companies and healthcare providers. The platform consists of three main components: web app, intelligent document processing (IDP), and the Generative AI Criteria Matching System.
The web app provides insurance and healthcare providers with a user-friendly interface, hosted on Amazon Simple Storage Service (Amazon S3) and served by Amazon CloudFront for fast and reliable access. User authentication is managed by Amazon Cognito, while the backend utilizes Amazon API Gateway, AWS Lambda, and Amazon DynamoDB for secure data storage and management.
Figure 1 – Product screens from the Treatline platform.
Intelligent Document Processing
IDP is a key component of the Treatline platform, responsible for extracting, collecting, and interpreting data from medical records and documents, as well as insurance claim forms.
By utilizing advanced techniques, such as natural language processing (NLP), optical character recognition (OCR), and computer vision, IDP ensures accurate and efficient data processing for both insurance and healthcare professionals. This significantly reduces the time spent on administrative tasks, allowing them to focus on patient care and claim processing.
Advanced Text and Structural Data Extraction
IDP leverages Amazon Textract to extract both text and structural information from medical and insurance documents, such as tables and key-value pairs. This powerful OCR technology enables the platform to recognize and process complex documents with a high degree of accuracy, ensuring all relevant information is captured and utilized in the prior authorization process.
Figure 2 – AWS reference architecture.
In-Depth Document Analysis
Once the text and structural data have been extracted, IDP utilizes Amazon Comprehend and Amazon Comprehend Medical to further analyze the content of medical and insurance documents. These AI-powered services identify critical information such as entities, medical terms, and relationships, allowing the platform to gain a deep understanding of each patient’s medical history, insurance coverage, and requirements.
To enable quick and efficient retrieval of the extracted information, IDP uses Amazon Kendra and Amazon CloudSearch to create smart search indexes. Amazon Kendra is a machine learning-based enterprise search service, while Amazon CloudSearch is a fully managed search service. Both services allow users to search and retrieve relevant medical and insurance data effortlessly, further enhancing the platform’s efficiency.
The extracted information is then indexed and stored in an Amazon S3 bucket, with metadata stored in a DynamoDB database. This enables efficient retrieval and analysis of the data when generating prior authorization requests and processing insurance claims.
Scalable and Efficient Document Processing
To handle high volumes of documents and ensure the platform remains responsive, IDP employs an asynchronous processing approach. This is managed through Amazon Simple Notification Service (Amazon SNS) and Amazon Simple Queue Service (Amazon SQS), which control the flow of documents awaiting processing.
This architecture ensures the platform can scale to meet the demands of healthcare and insurance providers, regardless of their size or the number of prior authorization requests and claims they submit.
Seamless Integration with Treatline Platform Components
IDP is fully integrated with the other components of the Treatline platform, such as the web app and the Generative AI Criteria Matching System. This seamless integration allows healthcare and insurance providers to benefit from the power of AI and advanced data processing techniques, streamlining the prior authorization process and improving efficiency across the board.
By combining the capabilities of IDP with the other components of the Treatline platform, healthcare and insurance providers can significantly reduce the time and effort required for prior authorization and claim processing, leading to better patient outcomes and more efficient healthcare and insurance delivery.
Enhanced Decision-Making with Generative AI
The Generative AI Criteria Matching System efficiently processes medical summaries and matches them with the appropriate criteria for specific procedures. Let’s consider an example.
Input Medical Summary: Patient is a 52-year-old female presenting with cervical radiculopathy and upper extremity weakness. The patient has a positive Hoffmann sign, disturbance with coordination, and no signs of hyperreflexia or positive Babinski sign/clonus. MRI results confirm cervical spondylosis causing spinal instability.
The patient has undergone seven weeks of conservative treatments, including a structured program of physical therapy, home exercise program prescribed by a physical therapist, and epidural steroid injections. The symptoms persist, and the patient’s quality of life is significantly impacted.
Output – Criteria Matched: Procedure Name: Anterior Cervical Decompression with Fusion (ACDF) – Single Level
CPT Code: 22548
- (YES) Cervical radiculopathy or myelopathy from ruptured disc, spondylosis, spinal instability, or deformity.
- (YES) Evidence of cervical spondylosis causing spinal instability.
- (YES) Persistent or recurrent symptoms/pain with functional limitations unresponsive to at least six weeks of appropriate conservative treatment.
- (YES) Documented failure of at least six consecutive weeks in the last six months of any two of the following physician-directed conservative treatments:
- (NO) Analgesics, steroids, and/or NSAIDs.
- (YES) Structured program of physical therapy.
- (YES) Structured home exercise program prescribed by a physical therapist, chiropractic provider, or physician.
- (YES) Epidural steroid injections and or selective nerve root block.
The Generative AI Criteria Matching System is a crucial part of the Treatline platform, designed to match medical summaries with both medical and insurance criteria. It ensures prior authorization requests and insurance claims are tailored to meet the specific requirements of each payer, improving the likelihood of approval and reducing the need for time-consuming peer-to-peer reviews.
Efficient Matching of Medical Summaries and Criteria
Treatline’s platform incorporates advanced technology to efficiently process medical summaries and match them with relevant criteria, ensuring comprehensive prior authorization requests and insurance claims.
Leveraging Amazon SageMaker JumpStart, a machine learning hub offering algorithms, models, and ML solutions to access the FLAN-T5 XXL model, the Criteria Matching System accurately identifies essential information in each medical summary and aligns it with the appropriate criteria. This advanced matching capability streamlines the prior authorization and claims processing workflows, resulting in faster approvals, reduced administrative burden, and improved patient care.
By utilizing Amazon SageMaker JumpStart, Treatline maximizes the benefits of the FLAN-T5 XXL model. This state-of-the-art model is specifically designed for natural language processing tasks, allowing for accurate understanding and generation of text.
The Criteria Matching System efficiently processes medical summaries, extracting crucial details such as allergies, diagnosis, measurement results, substance use, and treatment procedures. The system matches this information with the relevant criteria, ensuring each request meets the specific requirements of the payer.
By incorporating these technologies into the Treatline platform, healthcare and insurance providers can enjoy a streamlined and effective prior authorization and claim processing experience. The efficient matching of medical summaries and criteria tailored to meet payer requirements leads to faster approvals, reduced administrative burden, and enhanced patient care outcomes.
With Treatline’s advanced technology and powerful AWS services, healthcare providers can optimize their operations and deliver timely and appropriate care to patients.
The implementation of the Treatline platform has led to significant improvements in the prior authorization process for healthcare providers. These benefits are reflected in several key metrics, showcasing the platform’s effectiveness and efficiency.
Benefits for the Provider
- Cost savings and additional revenue: The Treatline platform reduces the administrative burden and streamlines the prior authorization process, leading to cost savings and the potential for additional revenue.
- Improved employee well-being: By automating routine tasks and reducing administrative workload, the platform helps prevent burnout among medical staff, promoting a healthier work environment.
- Faster A/R turnover: The efficient prior authorization process enabled by Treatline results in a quicker A/R payback period, improving the overall revenue cycle for healthcare providers.
Estimated Savings and Revenue Increase for the First Year
- 30% fewer peer-to-peer reviews: The Treatline platform’s intelligent criteria matching system significantly reduces the need for time-consuming peer-to-peer reviews, contributing to a more efficient prior authorization process.
- 70% less time spent on administration: The automation and AI capabilities of the Treatline platform minimize the time medical staff spend on administrative tasks, freeing them up to focus on patient care.
- Significant savings and profit increase: The implementation of the Treatline platform has the potential to result in substantial annual savings and a considerable increase in profit for healthcare providers.
By adopting the Treatline platform, healthcare providers can transform their prior authorization process, leading to improved efficiency, cost savings, and better patient outcomes. These results highlight the value of incorporating advanced automation and AI technologies into healthcare administration.
Neurons Lab is an innovative technology company specializing in AI/ML and software development. As a trusted partner in the healthcare and fintech industries, Neurons Lab is committed to delivering cutting-edge solutions that enhance the efficiency and effectiveness of processes in these sectors.
The collaboration between Neurons Lab and AWS ensures the Treatline platform benefits from the most up-to-date and reliable technology available, providing healthcare providers and insurance companies with a robust and dependable prior authorization solution.
Neurons Lab’s team of experts are well-versed in the unique challenges and requirements of the healthcare and fintech industries, as well as the potential benefits of incorporating AI/ML technologies.
Neurons Lab – AWS Partner Spotlight
Neurons Lab is an AWS Advanced Tier Services Partner and globally distributed AI R&D company that helps deep tech innovators to accelerate data-driven products development.