By Craig Mercure, Ciox, EVP Retrieval Solutions
It’s not as easy as you think. Even in today’s world of pervasive EMR adoption, digital connections, standards for data exchange, and increasing interoperability, the reality is that many organizations need access to rich clinical content from medical records that can only be obtained through rigorous and complex retrieval methodologies. These needs span across both private sector and Government stakeholders. The Federal Government has many needs and uses for big health data sets. The reasons why data is needed may seem basic, but it is worth reviewing them to understand the challenges and opportunities facing our industry today.
From Medicare and Medicaid risk adjustment to population health studies in the life sciences and more, Government entities behave much like health insurers, or payers, when it comes time to request medical records from providers. Often, their requests come in the form of massive, sweeping requests for large volumes of records in a process described as “batch retrieval.”
At the end of the month, the Government entity issues requests for a giant batch of records from a provider, and upon receipt of this aggregated request, the provider office sets to work collecting and delivering the medical records on behalf of the requestor.
The collection and delivery process becomes more complicated when scaled to larger medical groups or health systems. Most records are requested in order to aggregate rich clinical content to make population-level decisions based on the data available. Making decisions about an entire subgroup of, say, Medicare recipients, requires that the data about the group be flexible, combinable and interoperable.
Even today, most records kept by provider offices are either on paper to be faxed when requested or exist in a disconnected EMR format that cannot be combined with the records coming from other providers. For Federal agencies looking to connect the information and make better decisions at the population level, the batches or individual records delivered from providers pose a challenge.
Thanks to artificial intelligence (AI) technologies, it is possible to acquire and structure health records automatically, even for research and enterprise-grade data sets, and to break down the illiquidity that plagues the Healthcare industry and Government alike. AI, along with root cause analysis (RCA), robotic process automation (RPA) and optical character recognition (OCR), are used together to automate the delivery of more complete and interoperable pictures of batch data.
Natural Language Processing (NLP) can convert millions of records and unstructured notes into electronic documents with a high level of accuracy. These tools can recognize when to loop in human intervention to ensure greater accuracy or account for subjective references to key terms or attributes that could have clinical significance. Coding staff who review the digitization of these records only need to review tiny portions of the massive stores of data being batched, and manually recapture or adjust the percentage of data flagged for review. Their actions also systematically feed a machine learning and artificial intelligence feedback loop to continuously train the system from each manual intervention and improve accuracy in the automated process for future data extraction.
With the combination of the right technology and services, we can access centralized data that helps increase efficiencies for private organizations and Government agencies alike and, more importantly, offers profound insights designed to drive better patient care across the Healthcare spectrum.
About the Author
Craig Mercure, Executive Vice President of Clinical Data Acquisition & Insights, oversees business development, including strategic planning, sales, client services, marketing, product development, finance and communications. He also participates in the infrastructure and development of the company, which includes: systems, processes, pipeline management, trade support, marketing, facilities, personnel recruitment and development. Over the past 17 years, Craig has worked in executive leadership positions within the electronic medical record and medical documentation industry.