VA Wants to Automate Digitization of its 5-Mile-High Electronic Health Record Backlog
Veterans Affairs is looking to robotic process automation to help digitize its backlog and integrate with the new electronic health record platform.
The Veterans Affairs Department is on the cusp of deploying a massive electronic health records system and wants to know if robotic automation can help.
The VA originally planned to deploy the first instance of the new EHR system at the Mann-Grandstaff Medical Center in Spokane, Washington, this past March. However, officials announced in February that the initial rollout would be delayed, later telling Congress the agency planned to launch the new EHR system in July. That deadline has since been pushed further as the VA deals with the ongoing COVID-19 pandemic.
As the system reaches full operating capability, VA health care providers and staff will need to ensure all health records—old and new—are digitized and linked with the correct patients and clinicians. Integrating and deploying an electronic health record system is incredibly difficult, made more so if the data being digitized is unstructured or otherwise not machine-readable—a perennial problem with health data.
VA is looking at robotic process automation, or RPA, as a potential solution and issued a request for information Wednesday to hear from the contractor community.
“The Department of Veterans Affairs is investigating automation solutions to streamline the flow of external clinical document sets into its electronic health record system,” the RFI states.
In the near term, VA plans to use the automated system to cull some of its records digitization backlog, which, as of July 2018 had reached nearly 600,000 documents that, if stacked, would reach five miles high, according to the VA inspector general.
In the future, VA officials plan to integrate the RPA system with the current, agency-built VistA EHR system, as well as the new Cerner Millennium platform.
According to the RFI, a working solution will have three core capabilities:
- Ability to automate the document scanning process, including ingesting documents by “all methods used by external providers;” indexing the document in the EHR platform and connecting it to the right patient and clinician; correcting errors like missing data and upside-down or sideways page orientation; and identifying duplicate documents.
- Identify documents that cannot be ingested automatically and flag them for human processing.
- Provide metrics to remediate problems and help with future RPA planning.
The RFI offers additional details on specific deliverables.
Success will be measured by the number of documents processed within a given timeframe, including comparing that number with those that need human intervention; how accurately documents are linked to the correct patient and care provider; and the ability for the algorithm to learn and improve over time, especially when coversheets with relevant, well-marked information are not included with the documents.
Responses to the RFI—which includes three questions on VA’s specific use case and five general questions about the vendor’s automation solutions—are due by 12 p.m. July 30.