Overview of DRG Reviews
DRG reviews play a crucial role in ensuring accurate reimbursement for healthcare services. The two main types of DRG reviews, clinical DRG audits, and coding reviews bring different approaches to accurate reimbursement. While coding reviews focus on the accuracy of medical coding, clinical DRG audits take a broader approach by analyzing industry guidelines and medical records to determine the appropriate reimbursement for specific DRG concepts.
In healthcare, there are often competing guidelines for the same condition. For example, there are different guidelines for diagnosing and classifying sepsis, Sepsis 2, and Sepsis 3. These competing guidelines can create confusion and inconsistency in reimbursement decisions. By leveraging industry guidelines and white papers, clinical DRG audits ensure that the most up-to-date and relevant guidelines are applied, leading to fair and accurate reimbursement.
Building the Model
To build an effective clinical DRG audit model, you need a clinical review board consisting of expert coders and clinicians. This board analyzes industry guidelines, white papers, medical records, and claims data to identify relevant features associated with a particular DRG concept. These features include patient demographics, diagnoses, lab results, medications administered, procedures, and diagnostic test results.
By leveraging the outcomes of previous DRG audits, these features are analyzed to determine their weight of importance and incorporated into a decision tree. Additional clinical and coding rules are also incorporated into the algorithm to ensure adherence to coding guidelines and address time constraints and documentation requirements.
Feature engineering is a critical step in the model-building process, where AI technology utilizes various methods to extract the required data. For example, clinical natural language processing (cNLP) processes clinical notes and extracts features like diagnoses and impressions within radiology reports. Labeled clinical text is required for this process. Ontologies such as SNOMED are leveraged to identify the ICD-10 and CPT codes driving the DRG. Additionally, table detection and data analysis techniques are used for laboratory results and medications.
Once the model is built, it undergoes rigorous evaluation against many claims associated with the concept. The model’s output is evaluated by both coders and clinicians for accuracy. Based on their feedback, the model is refined by adjusting the algorithm, incorporating additional filters, and adding features until the desired accuracy threshold is met. This iterative process ensures the model performs well across different medical record formats and claims.
Deployment and Continuous Maintenance
After achieving satisfactory performance, the model is deployed into the production environment for the clinical team to leverage for that specific DRG concept. The model is presented in an intuitive interface that allows users to easily validate the output against the source medical record. Transparency is a key aspect of the deployment, as each feature incorporated into the model is presented back to the auditor, providing insight into the decision-making process. This allows customers to provide detailed feedback, improving the model’s accuracy and effectiveness.
Continuous maintenance is crucial to ensure the model remains accurate and up to date. The clinical review board continuously evaluates industry guideline changes, ensuring that the models remain accurate and aligned with the latest standards.
Benefits of Using AI Predictive Models
Utilizing a predictive model to support DRG reviews offers several benefits to healthcare organizations.
Efficiency and Time Savings
The average medical record can span hundreds of pages, making the audit process time-consuming and labor-intensive. However, a predictive model narrows the focus to key clinical indicators, significantly reducing the time and effort required for auditing. By automating the identification of relevant features, healthcare organizations can streamline their audit process and allocate resources more efficiently.
Streamlined Audit Process
By leveraging the APIs provided by predictive models, healthcare organizations can select which claims require closer attention. This targeted approach improves the findings rate for DRG concepts, ensuring that audits focus on cases most likely to have an impact. Furthermore, these models enable organizations to generate insights and analytics to detect patterns and trends across their provider network. This information is valuable and can be used to identify areas for improvement and optimize reimbursement strategies.
Improved Consistency
Auditors may interpret guidelines differently, leading to inconsistencies in reimbursement decisions. Additionally, some organizations may lack clear policies or guidelines for specific DRG concepts, resulting in further variations in reimbursement. By implementing a predictive model, organizations can standardize their approach to reviewing and auditing claims, increasing consistency and reducing provider abrasion. This standardized approach provides clarity and ensures fair and accurate reimbursement for healthcare services.
Moving from Post-Pay to Pre-Pay
Health plans strive to pay claims correctly the first time, but this requires efficient processes to ensure a quick turnaround to payment. Predictive models are designed to scale and handle large amounts of data, allowing organizations to quickly analyze and pay claims appropriately. By leveraging the predictive capabilities of these models, health plans can confidently transition from post-payment audits to pre-payment reviews, reducing payment delays and improving overall financial performance.
DRG audits are critical to healthcare reimbursement, ensuring accurate and fair payment for healthcare services. Clinical DRG audits, such as those powered by CAVO’s Predict model, provide a comprehensive and accurate approach to assessing reimbursement. By leveraging industry guidelines, medical records, and claims data, these models streamline the audit process, improve consistency, and enable organizations to make data-driven decisions. The benefits of utilizing prediction models, including efficiency, streamlined audits, improved consistency, and the increased ability to move from post-pay to pre-pay, make them invaluable tools for maximizing healthcare reimbursement. Embracing the power of artificial intelligence in DRG audits is key to unlocking the full potential of healthcare reimbursement processes.
How Advent Can Help
CAVO® is a revolutionary technology platform that supports various complex clinical claims review processes performed by health plans and services companies. CAVO empowers highly skilled clinical and coding resources with AI-driven functionality that shifts the focus from “low value” tasks and requirements such as document access, search, analysis, and determination support within medical records and other unstructured data to “high value” tasks and requirements such as validation and final determination. The technology enables clinical and coding reviewers to easily access and structure medical records, itemized bills, and additional clinical data efficiently, consistently, cost-effectively, and profitably.
The CAVO Predict model has been built around clinical DRG reviews, which provide a more comprehensive and accurate assessment of healthcare reimbursement. CAVO Predict also incorporates a user feedback loop, allowing users to adjust and modify the features presented. This feedback loop helps the model pick up on trends and patterns for features driving the model’s output.
With over one million completed case reviews, CAVO delivers technology proven to scale up your medical record reviews and increase your team’s productivity by 300%-500%.