Access to Clinical Data

The challenges of chart chasing and other common methods of accessing data

In our last blog post, we discussed the different types and formats of data and why having robust data is so important for payers. In this installment, we’ll outline the different ways payers typically resort to accessing data and the drawbacks of each method.

Some of the following may be considered sources of administrative data, so the quality of clinical data extracted often depends on the structure and format of how it’s shared.

Claims submission and payment

For payers, the most common source of accessible clinical data is captured in the electronic X12-837 or ANSI-837 claim for services format. These contain lots of data regarding the provider interaction, plus reference information about the practice and the patient. Depending on a payer’s unique payment operations, some clinical data may also be retained from the X12-997 response acknowledging receipt of the claim and acceptance for further processing. Payers can also pull data from the X12-835 transaction, which shows the line items of the claim that will be paid or denied, and several other formats to ensure the data is complete and accurate.

Of course, electronic health records are not yet universal in healthcare, so payers still receive several paper submissions in UB-04 formats that must be converted to digital to glean any clinical data of value. The diversity of product offerings from commercial payers and Medicare plans, plus the state-by-state approach to Medicaid billing, introduce further variation. 

Although claims data is easy for payers to access, the lag of payment operations compromises its timeliness, and the data may also lack valuable information like vital signs, test results, and changes in health status and outcomes over time. 

Paper chart chasing

Many pockets of healthcare still rely on paper records—especially small, rural hospitals and independent practices. Because payers may have difficulty accessing a provider’s EMR, they must rely on “chart chasing” to obtain paper copies or visit a provider on-site and manually extract the data they need. In a recent market study, Verinovum found examples of startup Medicare Advantage plans that sent case managers to visit network providers each quarter to gather data points for Star reporting requirements. 

This practice remains common to support a specific plan offering, as opposed to a more enterprise approach to cost effectively capturing the right clinical data. Its high costs tend to be bundled as part of payers’ administrative burden. Payers must assume that the data captured on the paper record is consistent and accurate, yet any gaps or errors are passed along in the new, digital format. 

Batched spreadsheet by cohort

Payers also commonly make direct requests of specific provider networks to collect and share specific data elements by targeted member populations, or cohorts. Typically, this is driven by contract terms and specific product offerings. It places the onus of capturing the data on the provider, who then pulls it from various EMRs, registries and care workflow to manually populate a spreadsheet outlining a limited dataset. It’s usually done weekly or monthly on an Excel spreadsheet that the payer must ingest into its own data workflow.

With dozens of networks, a portfolio of products and multiple formats, it’s easy to see how this becomes a labor-intensive process that detracts from the member encounter.

EMR login credentials

Predictably, payers sometimes encounter resistance when they ask providers to submit batched spreadsheets, so some have negotiated login credentials so they can access the provider’s EMR and pull data themselves. Usually, this is allowed for a defined population and gives the payer access to clinical data closer to the member experience.

This tends to shift the labor burden of pulling and integrating data feeds from EMRs from the provider to the payer, who typically accesses EMRs when internal resources are available for pulling the data, reformatting it and integrating it into the target workflow. This option tends to also have lag, becomes a de-facto batch event given labor availability, and requires configuration across one or more internal use cases.

Chart pulls

Health plans can do “chart chasing” or chart pulling in any number of ways, including sending case managers to do it, employing a third-party vendor or by using a chart abstraction service to pull data from paper records and PDFs. Sometimes payers have their own chart abstraction technology that links to digitized paper records and extracts select data elements of need, with the most advanced versions involving machine learning and converting unstructured text to a usable format.

Regardless of the approach, chart pulls combine heavy labor charges and transaction fees. There’s also a lag associated with the timing of the abstraction and duration of the integration. It tends to be iterative and requires local-level configuration of labor or methods to draw datasets. Finally, every time a vendor changes a system or workflow, the chart pull method must adapt as well.

Direct connection to EMR

The Health Information Technology for Economic and Clinical Health (HITECH) Act was part of the American Recovery and Reinvestment Act of 2009 and created incentives to make the exchange of EMR data more seamless and efficient to support care coordination and lower administrative costs. The exchange of clinical data is supported through various interoperable formats, the most common of which are: 

  • HL7 versions 2.x and 3.x. These introduced significant new efficiencies, but they also posed challenges including bastardized formats during EMR upgrades or customizations, inconsistent field definitions and quality control standards, and inconsistent data availability by EMR and HL7 message types.
  • Health Information Exchange (HIE) participation. While they hold great promise to improve the timeliness and robustness of data, HIE maturity and effectiveness varies widely across the country. Because even the most successful HIEs do not enjoy 100% penetration, payers must find another method to get data from all providers.
  • CMS required HL7 FHIR format. CMS required health plans to use the Fast Healthcare Interoperability Resources (FHIR) standard starting July 1, 2021, to make five types of data available to members to benefit patient engagement. By agreeing to FHIR and improved data standards upfront, the costs of clinical data collection and the efficiency of both clinical and administrative workflow can be greatly improved. However, most EMRs are not yet producing data in this format, so a payer may need an intermediary to receive HL7 2.x and 3.x data feeds, which can then be parsed and cleansed to resolve specific patient identity and encounter detail. From there, payers will likely need further data curation and enrichment to support specific use cases. Depending on the mix of formats utilized, further reformatting may be necessary so the payer can drop this data into an FHIR repository of their own. 

FHIR presents payers with an opportunity to get timely, robust clinical data at a relatively low cost, but it will require committing to an enterprise approach that sees them abandon a dependence on iterative and siloed approaches across claims data, spreadsheets, EMR logins and chart pulls. It will also require more concerted business and operational development around mechanisms that can leverage FHIR standards, including incenting providers to drive data to FHIR-enabled EMRs and HIEs. 

Ideally, both payers and providers will benefit from a more unified and dynamically updated view of all core clinical and administrative member data.