confidence scores enable targeted chart reviews

AI is no replacement for human judgement but it can go a long way towards simplifying and streamlining data management and analysis.

AI-enabled chart review helps providers and plans that coordinate care for complex populations work smarter, not harder, by shifting the focus from volume of charts targeted to precision targeting of charts.

One way AI makes this possible is with statistical probability and confidence scoring. This method typically assigns confidence scores which reflect the statistical probability that a chart review and/or follow up appointment will result in a confirmation that the suggestion is accurate. Coders or physicians can then review to determine if clinical documentation is sufficient to support the suggestion.

For providers, the ideal scenario is for support staff to vet suggestions with a high likelihood (at least 50%) and target outreach efforts and set appointments accordingly. Physician review can then be limited to the cases with the highest confidence scores (we recommend 90%) for an assessment and final decision. This should be incorporated into the workflow so that providers are prompted through the EMR to consider suggestions during encounters.

Plans can set a similar confidence score threshold depending on the size of their membership and scope of resources. By applying AI technology with statistical probability and confidence scoring capabilities, medical coders can focus on reviewing the highest value opportunities.  AI can also make each chart review more efficient by flagging sections and conditions that require review for quick and easy reference. This helps to streamline the audit process and frees up time for additional coding passes if needed before submission windows close. By targeting high priority charts, plans have a greater chance of capturing missed diagnoses that could significantly impact risk scores and preempting potential compliance issues that could result from overcoding and inadequate documentation.

shade