Millions of Medicare Advantage medical charts are retrieved and reviewed manually each year to improve accuracy and completeness of member health status. Traditional retrospective risk adjustment processes lack the technology and tools to predict and prioritize charts likely to support unreported diagnosis codes. The result has been a less efficient process, which can disrupt providers by requesting more medical charts than necessary. And that disruption can translate to provider abrasion.
It’s no wonder why there is provider abrasion. Optum conducted a study on how many of those charts retrieved from providers for risk adjustment actually supported unreported diagnosis codes. The study found that more than 50% of medical charts retrieved DO NOT support unreported diagnosis codes.
We can all relate to the mission of risk adjustment to drive better outcomes and improve the accuracy and completeness of member health status. How can you make that happen?
In the monthly Optum webinar series, “The Path to Risk and Quality Success,” a webinar was held entitled “Using AI to Improve Retrospective Risk Adjustment.” Kelly Tucker, vice president of risk adjustment operations, Optum, and Sam Diederich, vice president of strategy and business integration, Optum, discussed how applying artificial intelligence (AI) to your retrospective risk adjustment process can help to maximize outcomes through better coding efficiency, accuracy and completeness.
Is anyone using AI in their risk adjustment process?
We asked webinar participants if AI is powering any part of their risk adjustment process currently. Almost 30% of respondents are not using AI in any part of their risk adjustment process.
Results for using AI for chart targeting, coding or all parts of the process were relatively close at around 20%. Almost 5% of respondents are using AI for retrieval only.
How AI helps optimize the risk adjustment process
- Chart targeting — When we began using AI for chart targeting, we observed a reduction of 3% to 10% or more in chart retrieval requests. AI helps to precisely predict and prioritize charts likely to support unreported diagnosis codes. To create even greater accuracy, AI can be configured to exclude certain criteria.
How can you benefit? Besides the reduction in chart retrieval, provider abrasion can be reduced because less charts are requested.
- Chart retrieval — Without AI, most retrieval is done manually through on-site visits to providers, which disrupts the office environment or through other manual transfer methods. For AI-enabled chart retrieval, AI decides the method deemed most likely to be successful with providers. But how does AI decide? When analytics are aligned with provider preferences, that can facilitate even higher retrieval rates. In many instances, directly connecting to electronic health record (EHR) systems enables AI to target precisely which charts to retrieve and when.
How can you benefit? AI can help store provider preferences from prior retrieval seasons and also can often entirely remove the need for provider action from the workflow.
- Chart review — During the webinar, we talked about the manual process of transferring a chart image whether paper or from an EHR, and that a chart can be many, many pages. Imagine being the coder who just received a few large charts. The information they receive can be disjointed and disorganized. Now they must review that chart, find suspected conditions and confirm supporting documentation is present to decide what codes to assign. All of this is done, in most cases, with very little direction or orientation to the member and their specific disease profile.
AI evolves the chart review process so it is much more efficient. By employing AI to review medical charts before a human coder does, it is able to assist coders with the appropriate type of coding review that may lead to a more complete and accurate record. Data fed into AI can be trained to find specific diagnosis code suspects present in the medical record. It can also help the coder understand the member and their specific disease profile to help the coder code the condition accurately to the right level of specificity. By implementing this workflow into our own processes, Optum AI has enabled an 11-point improvement in our coding validation rates while also enabling more efficient chart review.
How can you benefit? AI helps the process to be more precise by capturing previously unreported diagnosis codes. In fact, in addition to increased validation rates, we also observed a 15% increase in conditions coded during chart review by leveraging proprietary AI-enabled chart review versus traditional methods.
Do you need to work with multiple vendors?
It can be a lot of work, hassle and expense to manage the vendors handling one or more steps of the retrospective risk adjustment process. During the webinar, we asked attendees how many vendors they use for their risk adjustment process. Half of the participants work with one vendor. Results for those working with two vendors or three or more were almost tied at a little over 24%.
A comprehensive process just makes more sense
A comprehensive retrospective risk adjustment solution enhanced with AI enables a smarter, highly efficient chart review process while helping to maintain coding accuracy and completeness.
It is important to view your retrospective program as an end-to-end workflow versus individual components. When you have a closed comprehensive retrospective risk adjustment system, AI capabilities and technology build on each other to optimize the process to help you get the best performance. The biggest advantage of a comprehensive solution is that AI is passed throughout the retrospective risk adjustment process. Having multiple vendors handling various components within the process limits the effectiveness of AI. In most instances, AI is constrained to only specific components of the workflow and is not continuously passed from one process to the next.
Start modernizing your process by decreasing the number of vendors that handle one or more components in your workflow. By reducing the number of vendors, you can streamline what should be one holistic end-to-end seamless program. You may also help reduce stress on overworked teams managing those vendors.
Key benefits of using AI to modernize your process
There are many benefits health plans may achieve from a transformed and modernized chart retrieval and review process powered by AI. Here are the top five:
- Maximize outcomes through better coding efficiency, accuracy and completeness — AI facilitates more precise capture of supported unreported diagnosis codes.
- Decrease need for multiple vendors and the associated management and expense.
- Deliver tangible, time-saving operational efficiencies to your provider network — precise prediction and prioritization enables less charts to be requested because AI can be configured to automatically exclude charts unlikely to contain supported unreported diagnosis codes before they are retrieved.
- Increase completeness of accuracy and codes found.
- Help reduce waste in the health care system.
On-demand and upcoming webinars
Please visit “The Path to Health Plan Success” webinar series page to explore upcoming topics, register for the next monthly webinar or watch one of our other webinars in this series on demand.
If you missed the webinar on “How AI Improves Retrospective Risk Adjustment” or one of our other webinars in this series, you may watch on demand.
To learn more about how AI can transform the retrospective risk adjustment process, visit our web page. While you’re there, download the infographic that shows at a glance how AI can enhance your risk adjustment process. To go deeper on why AI, download the white paper.
About the authors
Vice President of Risk Adjustment Operations, Optum
Kelly Tucker is responsible for strategy and solutions transformation for Optum Risk and Quality Solutions. She has over 15 years of health care experience, spending over a decade leading day-to-day operations for prospective and retrospective risk adjustment.
At Optum, Kelly and her current team are focused on risk adjustment modernization initiatives. Kelly believes in driving innovation to improve the customer experience for payers and patients of Medicare, Medicaid and commercial exchange programs. Prior to that, Kelly led operations overseeing 1,500 medical coders and the retrieval of over 7 million medical records annually.
Vice President of Strategy and Business Integration, Optum
Sam Diederich has spent nearly 10 years in the health care industry. He develops innovative strategies for payers in the areas of risk adjustment, utilization management and government-sponsored quality of care programs. He leads the growth, partnerships and integration strategy for the Optum quality and risk adjustment business. Sam is a graduate of the University of St. Thomas’ Opus College of Business, where he studied financial management and economics.