You can’t manage risk if you’re only seeing part of it. Unfortunately, health plans are often making risk adjustment decisions using only 20 to 30 percent of available data. This inevitably leads to inaccurate portrayal of patient acuity and unidentified risk. There are multiple issues that cause these shortcomings in identifying risk:
- Lack of a Plan – Many health plans haven’t systematically determined where to focus their medical chart retrieval and audit efforts. If resources are constrained and a submission deadline is approaching, plans should obtain charts that are most likely to contain documentation of uncompensated risk. From a compliance perspective, plans should also retrieve records for conditions previously submitted on claims that are common audit risks, such as acute HCCs.
- Incomplete Health Histories and Prioritizing Chart Reviews – Obtaining relevant medical records from healthcare providers is a labor-intensive process, and as charts from the provider network flow into the health plan, coders are unable to prioritize chart reviews based on the level of opportunity involved or proximity to submission deadlines. Additionally, coders may not have access to all of the members’ medical records in a consolidated view since records from various providers are collected separately, which may inhibit HCC identification and coding accuracy.
- Structured vs. Unstructured Data – Up to 80 percent of patient data is unstructured free text in medical charts of EHRs, meaning that it is not categorized in fields that can be easily analyzed using basic tools. This unstructured data requires manual review, which is error prone, time consuming, and ineffective when reviewing large volumes of data.
Strategies to Improve Risk Identification
Starting with simple initiatives that evolve into more complex strategies is an effective approach. First, health plans should identify the HCC-relevant diagnoses that have historically incurred high costs that are not reflected by the members’ risk scores. This creates a prioritized list of diagnoses that can be used in chart reviews to quickly flag high-risk, high-cost conditions, provided that they are documented in a compliant manner.
Furthermore, creating processes to better manage the handling and review of incoming medical charts is essential to improve risk adjustment efforts. It’s especially important to create processes that update member risk profiles as additional charts become available from different providers.
Prioritizing Chart Reviews
Relying on manual efforts to review medical charts for risk is a shot in the dark. Fortunately, there are tools to automate the process that use technology to analyze text within charts and provide coding suggestions for coders to review. A recent case study published by Health Fidelity details how this automation improved coder productivity by 4x, since it negated the need to manually review charts that contained no risk adjustment opportunity.
Analyzing Unstructured Data
Technology known as natural language processing (NLP) enables organizations to analyze unstructured clinical data – including admission notes, discharge summaries, progress notes, radiology/pathology/lab reports, nursing notes, psychiatric evaluations, etc. – to better identify risk factors. NLP analysis offers a 360-degree view of patient data, providing risk adjustment teams with access to all available information in a comprehensive workflow for complete and accurate risk identification. Health Fidelity’s NLP – the most advanced in healthcare – combines a proprietary library of clinical concepts, terms, and relationships with powerful processing technology to extract meaningful insights from unstructured clinical data and infer an accurate worklist to audit and manage risk adjustment.
With coordinated retrospective and prospective analytics powered by NLP as well as coder workflow optimization, organizations can gain greater control and transparency to successfully manage their risk adjustment programs.
There are numerous opportunities to improve risk adjustment programs, but they don’t need to be implemented all at once to realize benefits. An incremental approach – starting with HCC prioritization and gradually introducing automation – can yield substantial benefits that will grow as the complexity and effectiveness of programs evolve.