Clinical Document Automation For Healthcare: A Comprehensive Guide (+ Webinar Access)
5 Big challenges in automating healthcare documentation
During the 2023 annual convention of the Healthcare Information and Management Systems Society (HIMSS), Consensus Cloud Solutions conducted a series of interesting live experiments.
Each time a Consensus representative led one of the event’s sessions, they would ask a volunteer in the audience to fill out a standard patient form – the type of form we’re all handed on a clipboard and asked to complete before a doctor’s visit, asking for name, address, phone number, insurance information, etc.
The representative then fed that piece of paper into Consensus’s eFax solution, which converted the document to a PDF and then sent it to the AI-powered Consensus Clarity CD. The result? Seconds later, the audience saw the volunteer’s handwritten notes turned into a fully structured document with machine-readable data that could be ingested into any EHR.
This was just one real-world example cited during the recent Industry Dive webinar – Simplifying and Automating the Clinical Documentation Process – demonstrating how Natural Language Processing and other AI capabilities can help healthcare organizations dramatically reduce their clinical documentation burden.
1. We don’t have a universal patient ID.
Consensus’s Executive Vice President Bevey Miner describes this documentation burden as stemming primarily from two major inefficiencies in the healthcare industry.
As Bevey points out during the webinar, our patient identities are often scattered across the healthcare ecosystem. We might be represented with a member number for our insurance company, a patient number in our primary care practice’s database, an admissions number for a recent hospital stay, etc. Trying to match all of this information – and get it in a timely matter to the care teams that need it – is a major challenge because our medical data doesn’t exist in a single, comprehensive source.
2. We don’t have universally accepted data standards.
Here, Bevey notes that providers, payers, and other parties throughout the healthcare landscape use different tools, formats, and standards to communicate patient information – often time-sensitive or even urgent information that can’t be immediately ingested and acted on receiving it because the sender and recipient use different data standards.
How these challenges are undermining patients’ care and providers’ bottom line
The net effect of these challenges in healthcare document automation, the webinar panelists agree, is that exchanging patient information among providers almost always requires manual, labor-intensive, and error-prone processes.
For example, Ann Richardson, Founder of LeadWell Healthcare Consultants, notes that 80% of patient information arrives at providers’ practices as unstructured documents – such as faxes and even handwritten notes that computer systems can’t automatically read and process.
That creates the need for manual intervention – sometimes medical assistants, other times Registered Nurses – to review physical fax pages, scan their contents, and then manually re-enter key details such as patient demographic information into the EHR.
That takes clinicians off the hospital floor – and away from patients – to spend hours each shift on data entry. But even worse, Ann adds, is that this manual re-keying creates the risk of entering data into the wrong patient’s EHR record.
As Ann says during the webinar: “I’ve heard from patients whose care teams contact them through their patient EHR portal to suggest treatments for illnesses that the patient doesn’t even have – meaning the wrong documentation is making its way into patient records.”
Sachin Gangupantula, Co-founder and Head of Practice Operations at Valley Diabetes & Obesity adds that with so many new organizations entering the healthcare industry – including retail businesses such as Walmart and Amazon – it’s only becoming more difficult to coordinate information across all of these systems and keep patients’ records complete and accurate.
As Sachin points out, “Providers are still spending 60% to 70% of their time documenting.”
The answer: Streamline patient-data transfer with extractive AI
The panelists also agreed that the most accessible first-step solution to automating clinical documentation challenges is to apply proven tools for “extractive” AI – a branch of artificial intelligence that Ann is quick to point out helps with data automation without introducing the risks many people associate with artificial intelligence.
Suppose we can use AI to augment the transferring of patient data from one provider’s fax to another’s EHR. In that case, Ann explains, that’s a big win both for the healthcare organization’s efficiency and bottom line as well as the patient’s ability to receive better, more timely care.
“We don’t want to scare people with the sense that robots are going to replace humans here,” she says. With extractive AI – simply identifying data on an unstructured document like a fax page, and then placing that data in the appropriate fields in the patient record – “we can use automation to improve these data-automation processes and let the clinicians get back to where they can do the greatest amount of good: at their patients’ bedsides.”
The key takeaway from this webinar is the availability for healthcare organizations of all sizes – and even those with no in-house IT expertise – to implement a simple, customizable AI solution that can turn their high volume of inbound unstructured documents (handwritten notes, faxes, etc.) into machine-readable data that their EHRs and other systems can act on automatically.
Simplifying and Automating
the Clinical Documentation Process