Why Clinical Data is a Critical Aspect of Risk Adjustment

Adam600pxHealth plans are experts at handling claims data, but are far less adept at collecting and analyzing clinical data in patient charts. Claims data are readily available to health plans, but clinical data must be collected – often painstakingly – from providers, then stored and analyzed. Retrieved patient charts are typically reviewed manually by health plans or outsourced coders to identify diagnoses missing from claims for accurate reimbursement. Ideally, the review process could be streamlined to focus coders’ attention on high priority charts and improve coding accuracy. Clinical data in patient charts could also be used for other analytical purposes, such as prospectively identifying risk conditions that could impact a health plan’s bottom line if adequately documented.

Few health plans have all of these clinical data analysis capabilities. And those that do are likely to admit that their processes and technology have room for improvement.

 

Why Incorporate Clinical Data?

Clinical data contains the greatest details about patient health and risks that simply can’t be obtained from claims data alone. Additionally, only clinical data can validate the risk assumptions and predictions that are developed through the analysis of claims data.

So, why isn’t every health plan rushing to fully use clinical data in their risk adjustment programs? Because it’s difficult – very difficult.

 

The Obstacles

The first challenge is obtaining clinical data. Although health plans can request patient records from the providers that submit claims to them, the collection process is difficult. Currently, most patient record collection is done manually via fax, mail, or even by sending chart abstracters in person to provider offices. It’s a costly and time-consuming process.

Compounding the data collection challenges is the complex network of providers. The average Medicare patient sees six physicians per year. That means there are at least six charts to collect for a single patient per year – if the health plan wants to develop a longitudinal picture of health and risk.

The increased use of electronic health records (EHRs) is creating the potential for health information exchanges (HIEs), but its practical use in today’s industry is limited. Despite the rapid increase in EHR adoption, interoperability across disparate systems has yet to be realized. While the business case for HIEs is fairly straightforward for payers, getting providers to participate can be challenging. And questions around how HIEs are funded and how data exchange is priced still remain. However, it’s an important progression in the future of risk adjustment.

Subsequent challenges include handling the data – namely, its storage, integration, and retrieval/reuse. While health plans manage vast volumes of claims data, few have the infrastructure or processes to handle clinical data.

Similarly, the analytics needed to extract value from clinical data are simply not in use by the majority of health plans. This is especially true given that the majority of clinical data is unstructured free text that cannot be parsed without advanced technology. While health plans have sophisticated analytics in place for their claims and other administrative data, clinical data analytics (from a health plan perspective) are a relatively new frontier that hasn’t been readily explored.

 

A Perspective on the Future of Clinical Data Use

Health plans are highly experienced with technology infrastructures, processes, and productivity. There’s little doubt that they can, and will, leverage that experience and apply it to clinical data management in the future. That provides a fraction of the remedy for the challenge.

What’s missing are new data collection, analysis, and management capabilities. Technology advances such as process automation and virtual remote printing can be used to ingest chart data from provider systems in lieu of manually scanning or faxing. Then, optical character recognition (OCR) and natural language processing (NLP) can automate the analysis of large clinical data sets – regardless of each record’s native format – and provide meaningful insights, such as accurate coding suggestions and suspected care gaps. Advanced analytic capabilities powered by NLP have the potential to impact the health plan’s bottom line, improve population health management, and inform business decisions such as premium pricing, membership expansion, provider contracting, bid rate calculation, and more.

The Affordable Care Act is gradually reducing Medicare Advantage payment benchmarks. And CMS’s 2017 Final Notice, published on April 4, revealed that the new proposed hierarchical condition category (HCC) model would result in reduced risk factors for some of the most prevalent HCCs, further underscoring the importance of accurate documentation and coding. Traditional approaches to collecting and analyzing data used for risk adjustment are simply not enough.

By using technology to fully take advantage of clinical data, organizations can gain the greatest visibility into their risk and manage it appropriately.

Is your health plan ready?